Added

  • regscale client_runs commands: list shows the sync runs this machine left open, status shows one run, find names the run that holds a plan and source, and close closes a stuck run without mop-up
  • Every command that loads configuration now logs at INFO where its configuration came from and why (the working directory default, REGSCALE_CONFIG_FILE or --config), followed by a second line naming the environment variables that override it and how many secret-store references it holds. Values are never logged.
  • The compliance settings model now reads the vulnerability SLA mode and multi-agency flag from the platform.
  • Microsoft Defender and Defender for Cloud recommendations are now also reported as vulnerabilities, so they appear on the Vulnerability Status Board and in the FedRAMP VDR report, linked to their existing issues; this needs RegScale platform 6.36.2.0 or later, and older platforms keep creating issues only
  • Curated RSA Archer, Splunk, Microsoft Intune, ServiceNow, GitHub, GitLab, Sicura and Microsoft Sentinel commands in the Orchestration Hub command set (roh)

Changed

  • Trivy and Grype import_scans send each vulnerability once, with every asset that carries it, instead of once in every request that holds one of its findings, so a 750,000-finding import sends 20,000 vulnerability items (87 MiB) instead of 677,000 (1.2 GiB); --dry-run, --offset and --limit then count vulnerabilities, the closing log lines count them too and name the findings they cover, and groupScanFindings: false sends one item for each finding as before
  • A sync that finds another run of the same source still open for its plan now asks in a terminal whether to close that run and continue; answering y closes it without mop-up, which closes no records
  • Trivy and Grype import_scans send a folder's findings straight from the scan files instead of copying them into a multi-gigabyte staging file first, so a month of scans needs no large file on disk and no read of one from a slow shared drive; set stageScanFindings: true to stage them as before
  • Faster preparation of large flat-file scan imports (CSV, XLSX, JSON and XML), with no change to the data they create
  • A new init.yaml no longer carries issues blocks for aqua, defenderFile, ecr, nexpose, snyk, sonarcloud, veracode and xray, because the CLI never read them and RegScale takes issue due dates and status from the plan's compliance settings; blocks already in an existing init.yaml are ignored without a warning
  • The Nessus file import and the Tenable.sc sync now record one vulnerability per Tenable plugin, mapped to every affected asset, instead of one vulnerability per CVE. This follows Tenable's own definition of a finding (plugin, host, port and protocol), where the CVEs belong to the plugin and share one fix. The vulnerability's CVE field holds the plugin's first CVE, and its description lists every CVE of a multi-CVE plugin. On the first import after upgrading, the earlier per-CVE vulnerabilities, and the issues paired with them, close, and one vulnerability per plugin opens in their place.
  • The Trivy and Grype importers now stage their findings in small files that are removed when the import ends, instead of one file of up to several GB that stayed on disk, and they stage faster
  • AWS command help and the AWS docs now call the service the CLI reads "AWS Security Hub CSPM", matching the AWS rename; command names, options and IAM actions are unchanged
  • Warn at WARNING level when an SSP import falls back to a default impact level or categorization.
  • Scanner syncs now link a vulnerability to up to 50,000 affected assets by default instead of 10,000; an init.yaml that already sets maxAssetIdentifiersPerVulnerability keeps its value
  • Since 7.1.0, the Qualys VM scan import keys a host whose QG Host ID is a placeholder (0, none or unknown) on its IP address, so a host imported earlier under a placeholder ID gets a new asset keyed on its IP and its old asset stops receiving the host's findings
  • Every flat-file import in one import_all run stores Flat File Integration as the asset scanning tool, not the name of an import that ran before it
  • When Jira refuses an attachment upload, or a new ticket's attachment step fails in the ticketing sync, the log shows one error line and no stack trace
  • FedRAMP import-drf clears the saved operational requirement attachments of an OR deviation for an empty cell
  • regscale fedramp import-drf no longer sets a deviation's CVE from the Vulnerability Name column when the workbook has no CVE column
  • regscale tenable io sync_assets sends each MAC address in upper case split by :, so it rewrites a stored MAC Address that uses -
  • AWS findings sync no longer writes a Lambda console link that does not open the function, such as one for an AWS ISO region or for a function name that is not readable
  • AWS asset sync keeps an EFS file system's stored Disk Storage when AWS reports no usable size, instead of sending 0 or a negative or dropping the asset, and logs a warning for a reported size it cannot use
  • The next Wiz inventory sync moves images stored as Hardware to Software and keeps their stored OS version, and to keep images as Hardware you list VIRTUAL_MACHINE_IMAGE in wizHardwareAssetTypes with useWizHardwareAssetTypes on, which applies that whole list
  • Tenable and Qualys CIS results, Tenable STIG results, AWS Config failures, AWS Security Hub controls, Google Cloud misconfiguration findings and Wiz cloud configuration issues now show as misconfigurations, and AWS Inspector, Google Cloud vulnerability and Prisma Cloud advisory findings show as software vulnerabilities even without a CVE

Fixed

  • A JCAM profile value that cannot be mapped now fails and skips only that plan, in plan creation, front matter, agency-defined data and the dry run, and the import continues.
  • Failed file uploads now log a truncated, escaped response body instead of the raw text.
  • The TUI console runner now streams output lines longer than 64 KiB to completion and terminates the command if reading its output fails.
  • regscale jcam test_connection now exits 1 with a configuration message when jcamToken or jcamURL is unset or a placeholder, instead of reporting success without making a request.
  • JCAM import_parameters --learn-from now requires --crosswalk, so a learned mapping is never written into the installed package where an upgrade would discard it.
  • Corrected stale SORN/JOIN clearing so an empty value is no longer written to a required field, which stops the platform refusing the whole privacy save.
  • regscale jcam export_ssp now sends each authorization and front-matter value as plain text read from the field the mapping profile uses, so drift detection no longer flags identical values and --force no longer fails with a serialization error.
  • Fixed the JCAM SDK so it no longer retries POST or PATCH requests on 5xx responses or read timeouts, which could create duplicate POA&Ms; they now retry only on connect and pool errors and on 429.
  • A partly filled jcamSystemRoles block now stops regscale jcam import_ssp with exit code 1 and a clear message before any plan is written, instead of ending in a traceback part-way through.
  • JCAM export_poams now sends each POA&M's stored severity as the matching JCAM criticality instead of No Selection, and no longer writes that value into JCAM's severity weakness-type field.
  • JCAM exports now treat an unreadable drift-check response as drift and any accepted 2xx write, including an empty 204, as a success.
  • The JCAM POA&M export now skips a POA&M and counts it as failed when its milestones cannot be read, instead of sending it without them.
  • regscale jcam check_config now reports an unset jcamFrameworkCatalog 800-53r5 catalogue id, which import_controls needs, instead of passing on the shipped catalogue fallback.
  • The FedRAMP inventory workbook importer now checks the "Asset Type" column against the allowed asset types and stores "Other" with a warning for any other value, such as a hostname.
  • Wiz and Defender syncs, and the DevSecOps orchestrator, no longer swallow SIGTERM during best-effort fetches, so a stopped sync ends through its cleanup and closes its run.
  • GitHub repository sync now records a repository on its first run instead of failing until the second run.
  • Synqly asset syncs run with --filter no longer mark the assets the filter left out as Inactive.
  • Stop the FedRAMP POAM import from closing outdated findings when close_outdated_findings=False is passed.
  • Scanner and compliance syncs now log throttled progress lines when no terminal is attached (cron, jobs run-due, redirected output) instead of running silent.
  • A batch the server refuses is now logged once as a single line with its counts, and the traceback is kept at DEBUG, instead of printing full tracebacks (twice for vulnerabilities) for refused or retried batches.
  • A vulnerability that fails to queue is now recorded as a failed item in the run, so the run exits 3 under REGSCALE_STRICT_EXIT_CODES=1.
  • The FedRAMP POA&M import now exits non-zero under strict exit codes when rows fail to parse or issues are rejected, and one finding that raises no longer ends the import.
  • A compliance sync that cannot process one of its items now skips control assessments, issue mop-up and asset mop-up for that run instead of judging controls from the remaining items and closing the failed item's issue.
  • Flat-file imports now skip a row whose builder raises, log the file and row, withhold mop-up, and carry on with the rest of the file and folder.
  • A flat-file import with a missing security plan or component now logs one ERROR line and exits 1 before any file is fetched or imported, instead of logging a warning and exiting 0.
  • Synqly ticketing commands now record a failed ticket create, ticket link save or attachment upload in the exit code and log the RegScale issue and ticket involved.
  • Synqly ticketing sync now creates tickets by sending only the issue and its ticket fields, instead of every command option and the query filter.
  • Sentinel incremental syncs no longer close open issues and mark assets Inactive for incidents and alerts that did not change since the last run.
  • The AWS Inspector sync no longer closes findings through mop-up when a findings bucket failed partway through the pull, and the run is now recorded as a partial failure.
  • QRadar sync-events now disables mop-up when a pull reaches max_events, so findings cut off by the cap are no longer closed.
  • Jira issue creation: one refused create no longer ends the run, so the keys of the other created issues are still saved to RegScale and a failed create is counted as a partial failure.
  • Burp and Acunetix folder imports now run as one import, so a later file no longer closes the findings an earlier file reported.
  • Fixed regscale defender sync_acr_containers marking other registries' container images Inactive when it syncs one registry or cannot read one.
  • Incident sync to ServiceNow no longer creates a duplicate incident when the lookup of the linked incident fails, and the run now exits 1 when a lookup or the link write-back fails.
  • The sonarcloud sync_alerts command now closes RegScale issues that SonarCloud has resolved, and a run without --regscale_id reuses one assessment per project instead of copying every issue each run.
  • Made regscale aws sync_config_compliance record a failure, log a warning naming the API action, and skip the success message when AWS Config refuses access to a read the run needs.
  • Fixed regscale azure sync_intune --dry-run no longer creates or overwrites SBOM records.
  • Fixed regscale salesforce sync so one refused Salesforce case no longer stops the run or loses the case numbers of the cases already created, which caused duplicate cases on the next run
  • Hold Qualys server-side mop-up for a sync_qualys run whenever any reported finding is updated or skipped client-side, so findings Qualys still reports are no longer closed; because most re-syncs update or skip some findings, a finding Qualys has stopped reporting stays open until a run in which none is held back
  • A failed Tenable WAS export no longer leaves a partial cache that a rerun in the same hour syncs as a complete export.
  • A Client API batch request that fails with a timeout or a server error is now sent again as the same request, after waits of 5, 15 and 45 seconds, instead of falling back to sending its items one at a time, which was too slow to finish; when every send fails the sync stops and says to run it again
  • Trivy and Grype imports of large scan folders no longer hold every finding in memory until the run ends, so an import of 750,000 findings peaked near 600 MiB instead of 8.3 GiB in our tests
  • Trivy and Grype import_scans no longer stop with TypeError: 'digest_size' is an invalid keyword argument for openssl_blake2b() on hosts whose Python has no built-in blake2b module, such as some RHEL and UBI images
  • Databricks sync_assets and sync_findings no longer fail with "affected_hosts is ndarray, not a list" when reading findings straight from a query with --query
  • S3 prefixes for the Trivy and Grype imports now accept the wildcards , ? and [abc], so a prefix like trivy-scan finds its files again
  • Updated the container images to the patched Kerberos libraries
  • Updated a dependency of the HTML parser to clear a high severity finding in the container image scan
  • Logging calls in the FedRAMP import and config loading no longer raise logging errors, so the upload status code and stakeholder debug output are now logged.
  • A catalog update now names the control in the justification it records when it archives a control. It previously recorded the literal text {control_id}.
  • Group user lookup no longer fails with an AttributeError for groups that have users, and it adds each user's delegates once without repeating or looping on delegation cycles.
  • AWS inventory now logs a warning, not an error, when QLDB is enabled but the installed boto3 no longer includes the retired qldb service
  • jcam list_fields now sorts the fields in each tab by name.
  • jcam check_config now lists the destinations a tenant excluded with @drop, under a heading that marks them as chosen.
  • OSCAL profile import now reads the controls of every imports entry and every include-controls block instead of only the first.
  • OSCAL profile import now matches control ids regardless of case, zero padding and enhancement notation, and the log names each control it cannot find.
  • OSCAL catalog import now keeps the text of every guidance and overview part of a control instead of only the last one.
  • ControlObjective.from_dict no longer raises a TypeError and no longer turns missing values into the text "None".
  • A Qualys CIS benchmark row with no host IP no longer stops issue creation; the row adds no host and the control still gets its issue.
  • Queue a record created and then edited in the same run as a single create, so no id 0 update is sent that could fail a whole update batch.
  • The JCAM controls import now reports a value the plan's compliance settings do not offer once per run with a control count, instead of one warning per control that reprinted the plan's whole allowed-value list.
  • jcam check_config now ends with a one-line ready or not-ready verdict that lists only the items that block the import.
  • Fixed the GitLab findings sync failing on a group whose projects have no successful pipeline or security scan job yet; it now logs that no scan results are available and exits cleanly.
  • Fixed STIG checklist import so the comments and finding details from a checklist file are saved instead of always being empty; re-importing a checklist whose comments or details changed now updates the stored checklist, and a checklist action that raised a recursion error when an item was deleted now reports that it is not supported.
  • Qualys asset de-duplication no longer relies on a pluginId value that the Asset model silently dropped; existing assets are matched by their Qualys asset ID.
  • FedRAMP SSP import no longer crashes on objective and parameter ids in OSCAL control implementations.
  • Fixed a crash when mapping Synqly security findings that have no vulnerabilities, and DuroSuite group template variables that were silently dropped on group creation.
  • Refreshed the vendored platform schema snapshot and the generated model fields.
  • Creating the shared moka cache from several threads at once no longer builds more than one cache.
  • AWS, Microsoft Defender, Azure Container Registry and flat-file asset imports now leave an asset's public-facing flag unset when the vendor gave no exposure signal, instead of recording it as not public-facing.
  • Qualys container findings now send the real CVSS v3 vector and report every vulnerability in a container instead of only the first, so container vulnerability counts will rise.
  • Prisma Cloud, Microsoft Defender for Cloud, Qualys Total Cloud, Qualys CIS and Tanium now read their shipped issues.<integration> block from init.yaml (including noPastDueDates), and a fresh init.yaml now includes issues.noPastDueDates.
  • Import now keeps the original detection date and False Positive value (a No in that column is no longer read as a false positive) and restores the scheduled completion date the export shifted by one day (the import adds that day to the Scheduled Completion Date of every workbook, including one RegScale did not export, but leaves an unparseable date unshifted), and the Rev5 export lists every non-Closed POA&M on the Open sheet with its status shown as Status: <status> in the Comments column
  • Records the server rejects, and still rejects after the one retry, now count as failures, so REGSCALE_STRICT_EXIT_CODES=1 exits 3 when assets, vulnerabilities or issues fail to save.
  • regscale tenable io sync_compliance_controls no longer says that RegScale saved no control implementation when RegScale saved them and only reading them back failed; the run still exits 1 and says how many saved records got no assessment
  • The help for aws sync_iam and aws sync_kms now describes the assets each command writes: one AWS account asset for sync_iam, not one per user or role, and one asset per KMS key for sync_kms, both stored with asset type Other
  • When the batch request for an asset fails while another import saves the same plan, an asset import now tries the platform's single asset create or update, which never adds a second copy, and reports the asset as not saved if the platform does not have it or the other import holds the plan longer than the request can wait
  • Flat-file imports stop with a usage error, and leave the mapping file unchanged, when its mapping entry is 0, false, an empty string or an empty list instead of silently replacing it with the scan file's headers
  • The Qualys VM scan import now names each skipped host by its DNS name, then its NetBIOS name (the same order as asset names), then its FQDN, quoted and escaped in the skip warning, instead of reporting every host with only a NetBIOS name as one unnamed host
  • An Axonius v2 incremental sync that fails to write some records no longer moves its watermark past them, so the next sync retries them instead of reporting success with 0 records
  • If a component's protocol name, type or interconnection information value is empty, FedRAMP OSCAL SSP imports now store its port record, which RegScale rejected before
  • If a component port value or port range is missing or not valid, FedRAMP OSCAL SSP imports now log a warning and continue
  • FedRAMP OSCAL SSP imports now store every port range of a component protocol, not only the first one
  • If a Services, Used By or protocol cell is empty in a FedRAMP SSP Word document, the import now sends "Unknown"
  • Burp and Acunetix imports of a folder with two or more files no longer stop at the second file because another sync is still open
  • An AWS asset sync that retires a region's deleted assets itself (on a plan with assets of several regions, or when an inventory section could not be read) now takes a deleted "Off-Network" or "Decommissioned" asset out of the latest scan without changing its status, so the FedRAMP inventory export reads "No" for it in "In Latest Scan"
  • A line break or control character in an attachment name no longer starts a new line in the Jira or ticketing sync log
  • FedRAMP import-drf saves the operational requirement attachments of an OR deviation in their own field
  • regscale fedramp import-drf reads a deviation's CVE from the CVE column, so re-importing an exported Deviation Request workbook no longer replaces each CVE with its issue title
  • Fixed Wiz vulnerabilities sync keeping a resource's asset link Open after Wiz resolved the rule on that resource while the rule stayed open on another: the link now closes at the date Wiz resolved it there
  • regscale jira issues and regscale jira tasks (including --to-evidence) now work against Jira Data Center / Server: when Jira reports a Server or DataCenter deployment the issue search goes to /rest/api/2/search, paged on startAt, instead of Jira Cloud's /rest/api/3/search/jql, which Data Center rejects. A server that reports no deployment type is tried on v3 first and on v2 when it answers 404, 405 or 410. A failed search now ends with one line naming the HTTP status and the endpoint, without the query string, and for Data Center points at regscale jira_dc.
  • regscale axonius_v2 sync_saved_queries no longer prints a doubled period in the missing security plan error
  • regscale tenable io sync_assets now stores an asset that reports two or more MAC addresses. MAC Address holds one real address, the first in sorted order, so a new interface order does not change it. The command uses an all-zero or broadcast placeholder only for an asset with no other address. The asset's notes list the other addresses. If RegScale refuses an asset, the command now exits 1 instead of reporting success.
  • If Tenable.io reports an address as 0050.56a1.0f22 or 005056a10f22, regscale tenable io sync_assets now puts it in MAC Address
  • regscale tenable io sync_assets keeps an asset's MAC Address and notes the same when Tenable.io changes the letter case or separators of an address
  • Wiz compliance report issues for a failing control now count, list and name in their Asset Identifier only the resources that fail the control, and take their severity from the failing checks only
  • GCP findings now give their issues the Security Command Center recommendation as Recommended Actions instead of the generic sentence on RegScale 6.36.2.0 or later, including existing issues that still hold the generic sentence; a finding on the CISA KEV list can keep the KEV sentence, and an earlier platform ignores the recommendation
  • AWS Security Hub findings now count First Seen and the issue Due Date from the date the finding was first observed, with First Seen never later than Last Seen and the created or updated date used when a date is blank or not between 2000 and today. A consolidated vulnerability now takes the earliest First Seen and the earliest Date Created across the findings in its group, and the latest Last Seen, and counts its Due Date at the group's highest severity from that earliest First Seen, so an open issue Due Date can move earlier or later on the next sync and an issue whose new Due Date has passed becomes a POA&M
  • regscale aws sync_cloudtrail and regscale aws sync_s3 evidence records now list each control as passed or failed by the same rule as the run's control assessments, and name how many trails or buckets fail each control
  • AWS asset and findings sync now take a Lambda function asset's location from the region in its ARN, so a finding with no region no longer places a function in us-east-1 when its ARN names another region
  • AWS asset sync now fills in the URI of a Lambda function asset with its AWS console link, as AWS findings sync does
  • import_all run names each file it cannot read on one error line and stops with exit 1, where a traceback named only the first such file
  • AWS asset sync stores a DynamoDB table's Disk Storage in whole GiB instead of bytes, so a table of 2 GiB or more no longer fails the asset request; a table under 1 GiB shows 0
  • The AWS GuardDuty evidence record now says when the detector or findings read was incomplete, gives those counts as unknown instead of 0, and lists a control the read could not settle as not assessed instead of passed or failed
  • AWS GuardDuty sync logs each "not assessed" control warning and the malware findings warning once per run instead of repeating them
  • GCP sync_compliance and collect_evidence now log how many findings mapped to a control and how many had none
  • A list setting given as an environment variable, such as wizHardwareAssetTypes, now reads a JSON or comma-separated list instead of single characters, and a blank value falls back to the default
  • Wiz inventory and Wiz compliance and STIG asset stubs record a virtual machine image as Software, as the default hardware type list intends
  • A Wiz vulnerabilities sync no longer opens issues titled with the CVE twice ("CVE-2016-9267: CVE-2016-9267") for findings Wiz names by their CVE alone: the title now reads the CVE once, followed by the start of the description or, without one, the affected package; the next sync also retitles existing issues of this kind and updates their Security Checks text to match
  • The AWS command help and integration guide now match what sync_findings, the Audit Manager and AWS Config syncs and the asset sync do
  • regscale aws sync_assets no longer adds an Athena workgroup asset with no account number when AWS STS refuses the call and the security plan holds that workgroup for more than one account
  • The child jobs of a split Wiz inventory sync now read Wiz at the same time, so a later child waits only for the first child's remaining read and the writes ahead of it instead of every other child's full read
  • regscale aws sync_assets no longer marks Inactive an asset that regscale aws sync_findings_and_assets created for a resource the inventory does not list
  • regscale aws sync_assets --account_id now skips Athena workgroups, with a warning, when the AWS credentials belong to a different account, instead of saving that account's workgroups under the scoped account
  • aws sync_findings and aws sync_findings_and_assets with --account_id or --tags now close no vulnerabilities, and a run without these options closes the ones no longer reported
  • regscale aws sync_config_compliance no longer fails its compliance lookup or evidence collection when AWS Config returns a rule with no name, and its warning shows such a rule by its ID or ARN
  • regscale aws sync_config_compliance no longer closes the vulnerabilities and issues of passing controls when AWS Config returns a rule with no ARN or no name. The run names the rule in a warning. When the run checks that rule, it also leaves mop-up off and does not update control assessments. Fix or delete the rule in AWS Config to close records again
  • When AWS refuses part of the S3 bucket listing, regscale aws sync_s3 now warns and does not cache the partial list
  • regscale aws inventory collect now warns about each resource type that AWS refused to list
  • If AWS refuses the GuardDuty detector listing or a detector's details, regscale aws sync_assets now keeps the GuardDuty detector assets Active instead of marking them Inactive
  • When regscale aws sync_assets cannot read a resource listing, it now keeps those assets Active and names the type in a warning
  • regscale aws sync_s3 keeps its earlier evidence file when AWS refuses to return one of a bucket's settings or a bucket cannot be read
  • regscale aws sync_s3 now replaces the evidence file it attached earlier for the same AWS account instead of adding one more every run, and keeps files others added, files in an approval and files from a narrowed or partial run
  • regscale gcp sync_assets with --asset-types or --labels no longer sets the plan's other GCP assets Inactive
  • regscale gcp collect_evidence (and regscale gcp sync_findings --generate-evidence) now replaces the evidence file on each evidence record instead of adding one more file every run, and keeps files a person or another account added, files in an approval and files in an earlier version of the record
  • The gcp sync_compliance, collect_evidence and sync_findings --generate-evidence commands no longer set the plan's other compliance assets Inactive
  • The gcp sync_compliance command now closes its own earlier Fail of a control once no active finding fails it, and keeps any other Fail
  • Fixed gcp sync_findings with --sources or --severity-filter closing the plan's GCP findings that the filter left out; only an unfiltered run closes findings that are no longer in SCC
  • Fixed GCP findings sync keeping a vulnerability linked to an asset its finding no longer names: when a finding moves to another asset, the sync now maps the vulnerability to the new asset and closes its link to the old one, unless the vulnerability is not Open, when the old link stays Open
  • gcp sync_compliance and gcp collect_evidence now fail with exit 6 (4 when GCP refuses the credentials, 1 when the findings cannot be read) instead of ending successfully with nothing assessed when Security Command Center cannot be reached or its findings cannot be read
  • regscale jira issues no longer changes the "Due Date:" line of an issue description to ISO form (2027-02-04T07:19:41Z) on a re-sync; the line keeps 2027-02-04 07:19:41, or 2027-02-04 for a Jira due date
  • jira tasks now applies Jira's title, description and due date to a task in the same run that closes it; before, a task edited and moved to Done in Jira closed with its old values until a later run fetched the issue again
  • regscale jira issues no longer changes an issue that a scanner integration wrote, so a closed AWS or GCP issue stays closed
  • regscale wiz compliance_report now replaces the report file on the project's evidence record instead of adding one more file every run, and keeps files a person or another account added, reports in an approval and reports in an earlier version of the record
  • regscale wiz compliance_report and the deprecated regscale wiz sync_compliance now close the issue of a control that the latest report passes
  • A Wiz vulnerabilities sync now updates an end-of-life finding's vulnerability in place when Wiz renames the finding, instead of adding a second vulnerability for the same finding
  • A repeat OCSF ingest no longer creates a second asset for a named resource that has no uid, and it updates only a same-named asset whose notes hold that resource
  • A repeat OCSF ingest keeps the asset status and other asset values a person set
  • Wiz commands that get a query answer with errors and no data now print one error line that names the query and the errors Wiz reported, instead of several tracebacks, and optional steps that skip on such an answer (asset enrichment, the representative-asset collapse) say what they skipped without a traceback
  • Vulnerabilities that an integration builds itself (the Qualys VM report and CIS imports) now carry a finding type like the rest: software vulnerability when there is a CVE, otherwise other.
  • SARIF, Qualys Total Cloud, OpenText WebInspect and DevSecOps SARIF and Wiz CLI imports, and Trivy or Grype imports whose files share no scan day, no longer send the CLI machine's local date as the scan history date when none was given, so RegScale dates the scan history by the import time
  • regscale gcp sync_compliance with CIS_GCP, PCI-DSS or SOC2 no longer fails controls from Security Command Center tags of another standard or version (CIS Controls, CIS benchmarks other than 2.0, PCI DSS 4.0, SOC 1)
  • The FedRAMP inventory workbook importer now stores a common "Asset Type" such as "Server" as "Physical Server" (the shared asset type aliases) instead of "Other", or as "Virtual Machine (VM)" when the row's Virtual column is Yes; values that match nothing are still stored as "Other" with a warning.
  • Microsoft Defender asset sync now marks public IP address resources as public-facing, as PS-929 specifies, instead of leaving the flag unset.
  • A Nessus or Tenable.sc plugin that lists several CVEs now takes its KEV-listed CVE (the one with the earliest KEV due date) as the vulnerability's CVE, so the issue's KEV list, the KEV due date and the VDR KEV clock still apply when the KEV CVE is not the plugin's first. A plugin with no KEV CVE keeps its first CVE, and the description still lists every CVE. The vulnerability is still keyed on the plugin ID, and the issue's identity stays on the plugin's first CVE when the primary CVE is swapped, so a CVE joining the KEV catalog does not reopen the issue.

Added

  • regscale client_runs commands: list shows the sync runs this machine left open, status shows one run, find names the run that holds a plan and source, and close closes a stuck run without mop-up
  • Every command that loads configuration now logs at INFO where its configuration came from and why (the working directory default, REGSCALE_CONFIG_FILE or --config), followed by a second line naming the environment variables that override it and how many secret-store references it holds. Values are never logged.
  • The compliance settings model now reads the vulnerability SLA mode and multi-agency flag from the platform.
  • Microsoft Defender and Defender for Cloud recommendations are now also reported as vulnerabilities, so they appear on the Vulnerability Status Board and in the FedRAMP VDR report, linked to their existing issues; this needs RegScale platform 6.36.2.0 or later, and older platforms keep creating issues only
  • Curated RSA Archer, Splunk, Microsoft Intune, ServiceNow, GitHub, GitLab, Sicura and Microsoft Sentinel commands in the Orchestration Hub command set (roh)

Changed

  • Trivy and Grype import_scans send each vulnerability once, with every asset that carries it, instead of once in every request that holds one of its findings, so a 750,000-finding import sends 20,000 vulnerability items (87 MiB) instead of 677,000 (1.2 GiB); --dry-run, --offset and --limit then count vulnerabilities, the closing log lines count them too and name the findings they cover, and groupScanFindings: false sends one item for each finding as before
  • A sync that finds another run of the same source still open for its plan now asks in a terminal whether to close that run and continue; answering y closes it without mop-up, which closes no records
  • Trivy and Grype import_scans send a folder's findings straight from the scan files instead of copying them into a multi-gigabyte staging file first, so a month of scans needs no large file on disk and no read of one from a slow shared drive; set stageScanFindings: true to stage them as before
  • Faster preparation of large flat-file scan imports (CSV, XLSX, JSON and XML), with no change to the data they create
  • A new init.yaml no longer carries issues blocks for aqua, defenderFile, ecr, nexpose, snyk, sonarcloud, veracode and xray, because the CLI never read them and RegScale takes issue due dates and status from the plan's compliance settings; blocks already in an existing init.yaml are ignored without a warning
  • The Nessus file import and the Tenable.sc sync now record one vulnerability per Tenable plugin, mapped to every affected asset, instead of one vulnerability per CVE. This follows Tenable's own definition of a finding (plugin, host, port and protocol), where the CVEs belong to the plugin and share one fix. The vulnerability's CVE field holds the plugin's first CVE, and its description lists every CVE of a multi-CVE plugin. On the first import after upgrading, the earlier per-CVE vulnerabilities, and the issues paired with them, close, and one vulnerability per plugin opens in their place.
  • The Trivy and Grype importers now stage their findings in small files that are removed when the import ends, instead of one file of up to several GB that stayed on disk, and they stage faster
  • AWS command help and the AWS docs now call the service the CLI reads "AWS Security Hub CSPM", matching the AWS rename; command names, options and IAM actions are unchanged
  • Warn at WARNING level when an SSP import falls back to a default impact level or categorization.
  • Scanner syncs now link a vulnerability to up to 50,000 affected assets by default instead of 10,000; an init.yaml that already sets maxAssetIdentifiersPerVulnerability keeps its value
  • Since 7.1.0, the Qualys VM scan import keys a host whose QG Host ID is a placeholder (0, none or unknown) on its IP address, so a host imported earlier under a placeholder ID gets a new asset keyed on its IP and its old asset stops receiving the host's findings
  • Every flat-file import in one import_all run stores Flat File Integration as the asset scanning tool, not the name of an import that ran before it
  • When Jira refuses an attachment upload, or a new ticket's attachment step fails in the ticketing sync, the log shows one error line and no stack trace
  • FedRAMP import-drf clears the saved operational requirement attachments of an OR deviation for an empty cell
  • regscale fedramp import-drf no longer sets a deviation's CVE from the Vulnerability Name column when the workbook has no CVE column
  • regscale tenable io sync_assets sends each MAC address in upper case split by :, so it rewrites a stored MAC Address that uses -
  • AWS findings sync no longer writes a Lambda console link that does not open the function, such as one for an AWS ISO region or for a function name that is not readable
  • AWS asset sync keeps an EFS file system's stored Disk Storage when AWS reports no usable size, instead of sending 0 or a negative or dropping the asset, and logs a warning for a reported size it cannot use
  • The next Wiz inventory sync moves images stored as Hardware to Software and keeps their stored OS version, and to keep images as Hardware you list VIRTUAL_MACHINE_IMAGE in wizHardwareAssetTypes with useWizHardwareAssetTypes on, which applies that whole list
  • Tenable and Qualys CIS results, Tenable STIG results, AWS Config failures, AWS Security Hub controls, Google Cloud misconfiguration findings and Wiz cloud configuration issues now show as misconfigurations, and AWS Inspector, Google Cloud vulnerability and Prisma Cloud advisory findings show as software vulnerabilities even without a CVE

Fixed

  • A JCAM profile value that cannot be mapped now fails and skips only that plan, in plan creation, front matter, agency-defined data and the dry run, and the import continues.
  • Failed file uploads now log a truncated, escaped response body instead of the raw text.
  • The TUI console runner now streams output lines longer than 64 KiB to completion and terminates the command if reading its output fails.
  • regscale jcam test_connection now exits 1 with a configuration message when jcamToken or jcamURL is unset or a placeholder, instead of reporting success without making a request.
  • JCAM import_parameters --learn-from now requires --crosswalk, so a learned mapping is never written into the installed package where an upgrade would discard it.
  • Corrected stale SORN/JOIN clearing so an empty value is no longer written to a required field, which stops the platform refusing the whole privacy save.
  • regscale jcam export_ssp now sends each authorization and front-matter value as plain text read from the field the mapping profile uses, so drift detection no longer flags identical values and --force no longer fails with a serialization error.
  • Fixed the JCAM SDK so it no longer retries POST or PATCH requests on 5xx responses or read timeouts, which could create duplicate POA&Ms; they now retry only on connect and pool errors and on 429.
  • A partly filled jcamSystemRoles block now stops regscale jcam import_ssp with exit code 1 and a clear message before any plan is written, instead of ending in a traceback part-way through.
  • JCAM export_poams now sends each POA&M's stored severity as the matching JCAM criticality instead of No Selection, and no longer writes that value into JCAM's severity weakness-type field.
  • JCAM exports now treat an unreadable drift-check response as drift and any accepted 2xx write, including an empty 204, as a success.
  • The JCAM POA&M export now skips a POA&M and counts it as failed when its milestones cannot be read, instead of sending it without them.
  • regscale jcam check_config now reports an unset jcamFrameworkCatalog 800-53r5 catalogue id, which import_controls needs, instead of passing on the shipped catalogue fallback.
  • The FedRAMP inventory workbook importer now checks the "Asset Type" column against the allowed asset types and stores "Other" with a warning for any other value, such as a hostname.
  • Wiz and Defender syncs, and the DevSecOps orchestrator, no longer swallow SIGTERM during best-effort fetches, so a stopped sync ends through its cleanup and closes its run.
  • GitHub repository sync now records a repository on its first run instead of failing until the second run.
  • Synqly asset syncs run with --filter no longer mark the assets the filter left out as Inactive.
  • Stop the FedRAMP POAM import from closing outdated findings when close_outdated_findings=False is passed.
  • Scanner and compliance syncs now log throttled progress lines when no terminal is attached (cron, jobs run-due, redirected output) instead of running silent.
  • A batch the server refuses is now logged once as a single line with its counts, and the traceback is kept at DEBUG, instead of printing full tracebacks (twice for vulnerabilities) for refused or retried batches.
  • A vulnerability that fails to queue is now recorded as a failed item in the run, so the run exits 3 under REGSCALE_STRICT_EXIT_CODES=1.
  • The FedRAMP POA&M import now exits non-zero under strict exit codes when rows fail to parse or issues are rejected, and one finding that raises no longer ends the import.
  • A compliance sync that cannot process one of its items now skips control assessments, issue mop-up and asset mop-up for that run instead of judging controls from the remaining items and closing the failed item's issue.
  • Flat-file imports now skip a row whose builder raises, log the file and row, withhold mop-up, and carry on with the rest of the file and folder.
  • A flat-file import with a missing security plan or component now logs one ERROR line and exits 1 before any file is fetched or imported, instead of logging a warning and exiting 0.
  • Synqly ticketing commands now record a failed ticket create, ticket link save or attachment upload in the exit code and log the RegScale issue and ticket involved.
  • Synqly ticketing sync now creates tickets by sending only the issue and its ticket fields, instead of every command option and the query filter.
  • Sentinel incremental syncs no longer close open issues and mark assets Inactive for incidents and alerts that did not change since the last run.
  • The AWS Inspector sync no longer closes findings through mop-up when a findings bucket failed partway through the pull, and the run is now recorded as a partial failure.
  • QRadar sync-events now disables mop-up when a pull reaches max_events, so findings cut off by the cap are no longer closed.
  • Jira issue creation: one refused create no longer ends the run, so the keys of the other created issues are still saved to RegScale and a failed create is counted as a partial failure.
  • Burp and Acunetix folder imports now run as one import, so a later file no longer closes the findings an earlier file reported.
  • Fixed regscale defender sync_acr_containers marking other registries' container images Inactive when it syncs one registry or cannot read one.
  • Incident sync to ServiceNow no longer creates a duplicate incident when the lookup of the linked incident fails, and the run now exits 1 when a lookup or the link write-back fails.
  • The sonarcloud sync_alerts command now closes RegScale issues that SonarCloud has resolved, and a run without --regscale_id reuses one assessment per project instead of copying every issue each run.
  • Made regscale aws sync_config_compliance record a failure, log a warning naming the API action, and skip the success message when AWS Config refuses access to a read the run needs.
  • Fixed regscale azure sync_intune --dry-run no longer creates or overwrites SBOM records.
  • Fixed regscale salesforce sync so one refused Salesforce case no longer stops the run or loses the case numbers of the cases already created, which caused duplicate cases on the next run
  • Hold Qualys server-side mop-up for a sync_qualys run whenever any reported finding is updated or skipped client-side, so findings Qualys still reports are no longer closed; because most re-syncs update or skip some findings, a finding Qualys has stopped reporting stays open until a run in which none is held back
  • A failed Tenable WAS export no longer leaves a partial cache that a rerun in the same hour syncs as a complete export.
  • A Client API batch request that fails with a timeout or a server error is now sent again as the same request, after waits of 5, 15 and 45 seconds, instead of falling back to sending its items one at a time, which was too slow to finish; when every send fails the sync stops and says to run it again
  • Trivy and Grype imports of large scan folders no longer hold every finding in memory until the run ends, so an import of 750,000 findings peaked near 600 MiB instead of 8.3 GiB in our tests
  • Trivy and Grype import_scans no longer stop with TypeError: 'digest_size' is an invalid keyword argument for openssl_blake2b() on hosts whose Python has no built-in blake2b module, such as some RHEL and UBI images
  • Databricks sync_assets and sync_findings no longer fail with "affected_hosts is ndarray, not a list" when reading findings straight from a query with --query
  • S3 prefixes for the Trivy and Grype imports now accept the wildcards , ? and [abc], so a prefix like trivy-scan finds its files again
  • Updated the container images to the patched Kerberos libraries
  • Updated a dependency of the HTML parser to clear a high severity finding in the container image scan
  • Logging calls in the FedRAMP import and config loading no longer raise logging errors, so the upload status code and stakeholder debug output are now logged.
  • A catalog update now names the control in the justification it records when it archives a control. It previously recorded the literal text {control_id}.
  • Group user lookup no longer fails with an AttributeError for groups that have users, and it adds each user's delegates once without repeating or looping on delegation cycles.
  • AWS inventory now logs a warning, not an error, when QLDB is enabled but the installed boto3 no longer includes the retired qldb service
  • jcam list_fields now sorts the fields in each tab by name.
  • jcam check_config now lists the destinations a tenant excluded with @drop, under a heading that marks them as chosen.
  • OSCAL profile import now reads the controls of every imports entry and every include-controls block instead of only the first.
  • OSCAL profile import now matches control ids regardless of case, zero padding and enhancement notation, and the log names each control it cannot find.
  • OSCAL catalog import now keeps the text of every guidance and overview part of a control instead of only the last one.
  • ControlObjective.from_dict no longer raises a TypeError and no longer turns missing values into the text "None".
  • A Qualys CIS benchmark row with no host IP no longer stops issue creation; the row adds no host and the control still gets its issue.
  • Queue a record created and then edited in the same run as a single create, so no id 0 update is sent that could fail a whole update batch.
  • The JCAM controls import now reports a value the plan's compliance settings do not offer once per run with a control count, instead of one warning per control that reprinted the plan's whole allowed-value list.
  • jcam check_config now ends with a one-line ready or not-ready verdict that lists only the items that block the import.
  • Fixed the GitLab findings sync failing on a group whose projects have no successful pipeline or security scan job yet; it now logs that no scan results are available and exits cleanly.
  • Fixed STIG checklist import so the comments and finding details from a checklist file are saved instead of always being empty; re-importing a checklist whose comments or details changed now updates the stored checklist, and a checklist action that raised a recursion error when an item was deleted now reports that it is not supported.
  • Qualys asset de-duplication no longer relies on a pluginId value that the Asset model silently dropped; existing assets are matched by their Qualys asset ID.
  • FedRAMP SSP import no longer crashes on objective and parameter ids in OSCAL control implementations.
  • Fixed a crash when mapping Synqly security findings that have no vulnerabilities, and DuroSuite group template variables that were silently dropped on group creation.
  • Refreshed the vendored platform schema snapshot and the generated model fields.
  • Creating the shared moka cache from several threads at once no longer builds more than one cache.
  • AWS, Microsoft Defender, Azure Container Registry and flat-file asset imports now leave an asset's public-facing flag unset when the vendor gave no exposure signal, instead of recording it as not public-facing.
  • Qualys container findings now send the real CVSS v3 vector and report every vulnerability in a container instead of only the first, so container vulnerability counts will rise.
  • Prisma Cloud, Microsoft Defender for Cloud, Qualys Total Cloud, Qualys CIS and Tanium now read their shipped issues.<integration> block from init.yaml (including noPastDueDates), and a fresh init.yaml now includes issues.noPastDueDates.
  • Import now keeps the original detection date and False Positive value (a No in that column is no longer read as a false positive) and restores the scheduled completion date the export shifted by one day (the import adds that day to the Scheduled Completion Date of every workbook, including one RegScale did not export, but leaves an unparseable date unshifted), and the Rev5 export lists every non-Closed POA&M on the Open sheet with its status shown as Status: <status> in the Comments column
  • Records the server rejects, and still rejects after the one retry, now count as failures, so REGSCALE_STRICT_EXIT_CODES=1 exits 3 when assets, vulnerabilities or issues fail to save.
  • regscale tenable io sync_compliance_controls no longer says that RegScale saved no control implementation when RegScale saved them and only reading them back failed; the run still exits 1 and says how many saved records got no assessment
  • The help for aws sync_iam and aws sync_kms now describes the assets each command writes: one AWS account asset for sync_iam, not one per user or role, and one asset per KMS key for sync_kms, both stored with asset type Other
  • When the batch request for an asset fails while another import saves the same plan, an asset import now tries the platform's single asset create or update, which never adds a second copy, and reports the asset as not saved if the platform does not have it or the other import holds the plan longer than the request can wait
  • Flat-file imports stop with a usage error, and leave the mapping file unchanged, when its mapping entry is 0, false, an empty string or an empty list instead of silently replacing it with the scan file's headers
  • The Qualys VM scan import now names each skipped host by its DNS name, then its NetBIOS name (the same order as asset names), then its FQDN, quoted and escaped in the skip warning, instead of reporting every host with only a NetBIOS name as one unnamed host
  • An Axonius v2 incremental sync that fails to write some records no longer moves its watermark past them, so the next sync retries them instead of reporting success with 0 records
  • If a component's protocol name, type or interconnection information value is empty, FedRAMP OSCAL SSP imports now store its port record, which RegScale rejected before
  • If a component port value or port range is missing or not valid, FedRAMP OSCAL SSP imports now log a warning and continue
  • FedRAMP OSCAL SSP imports now store every port range of a component protocol, not only the first one
  • If a Services, Used By or protocol cell is empty in a FedRAMP SSP Word document, the import now sends "Unknown"
  • Burp and Acunetix imports of a folder with two or more files no longer stop at the second file because another sync is still open
  • An AWS asset sync that retires a region's deleted assets itself (on a plan with assets of several regions, or when an inventory section could not be read) now takes a deleted "Off-Network" or "Decommissioned" asset out of the latest scan without changing its status, so the FedRAMP inventory export reads "No" for it in "In Latest Scan"
  • A line break or control character in an attachment name no longer starts a new line in the Jira or ticketing sync log
  • FedRAMP import-drf saves the operational requirement attachments of an OR deviation in their own field
  • regscale fedramp import-drf reads a deviation's CVE from the CVE column, so re-importing an exported Deviation Request workbook no longer replaces each CVE with its issue title
  • Fixed Wiz vulnerabilities sync keeping a resource's asset link Open after Wiz resolved the rule on that resource while the rule stayed open on another: the link now closes at the date Wiz resolved it there
  • regscale jira issues and regscale jira tasks (including --to-evidence) now work against Jira Data Center / Server: when Jira reports a Server or DataCenter deployment the issue search goes to /rest/api/2/search, paged on startAt, instead of Jira Cloud's /rest/api/3/search/jql, which Data Center rejects. A server that reports no deployment type is tried on v3 first and on v2 when it answers 404, 405 or 410. A failed search now ends with one line naming the HTTP status and the endpoint, without the query string, and for Data Center points at regscale jira_dc.
  • regscale axonius_v2 sync_saved_queries no longer prints a doubled period in the missing security plan error
  • regscale tenable io sync_assets now stores an asset that reports two or more MAC addresses. MAC Address holds one real address, the first in sorted order, so a new interface order does not change it. The command uses an all-zero or broadcast placeholder only for an asset with no other address. The asset's notes list the other addresses. If RegScale refuses an asset, the command now exits 1 instead of reporting success.
  • If Tenable.io reports an address as 0050.56a1.0f22 or 005056a10f22, regscale tenable io sync_assets now puts it in MAC Address
  • regscale tenable io sync_assets keeps an asset's MAC Address and notes the same when Tenable.io changes the letter case or separators of an address
  • Wiz compliance report issues for a failing control now count, list and name in their Asset Identifier only the resources that fail the control, and take their severity from the failing checks only
  • GCP findings now give their issues the Security Command Center recommendation as Recommended Actions instead of the generic sentence on RegScale 6.36.2.0 or later, including existing issues that still hold the generic sentence; a finding on the CISA KEV list can keep the KEV sentence, and an earlier platform ignores the recommendation
  • AWS Security Hub findings now count First Seen and the issue Due Date from the date the finding was first observed, with First Seen never later than Last Seen and the created or updated date used when a date is blank or not between 2000 and today. A consolidated vulnerability now takes the earliest First Seen and the earliest Date Created across the findings in its group, and the latest Last Seen, and counts its Due Date at the group's highest severity from that earliest First Seen, so an open issue Due Date can move earlier or later on the next sync and an issue whose new Due Date has passed becomes a POA&M
  • regscale aws sync_cloudtrail and regscale aws sync_s3 evidence records now list each control as passed or failed by the same rule as the run's control assessments, and name how many trails or buckets fail each control
  • AWS asset and findings sync now take a Lambda function asset's location from the region in its ARN, so a finding with no region no longer places a function in us-east-1 when its ARN names another region
  • AWS asset sync now fills in the URI of a Lambda function asset with its AWS console link, as AWS findings sync does
  • import_all run names each file it cannot read on one error line and stops with exit 1, where a traceback named only the first such file
  • AWS asset sync stores a DynamoDB table's Disk Storage in whole GiB instead of bytes, so a table of 2 GiB or more no longer fails the asset request; a table under 1 GiB shows 0
  • The AWS GuardDuty evidence record now says when the detector or findings read was incomplete, gives those counts as unknown instead of 0, and lists a control the read could not settle as not assessed instead of passed or failed
  • AWS GuardDuty sync logs each "not assessed" control warning and the malware findings warning once per run instead of repeating them
  • GCP sync_compliance and collect_evidence now log how many findings mapped to a control and how many had none
  • A list setting given as an environment variable, such as wizHardwareAssetTypes, now reads a JSON or comma-separated list instead of single characters, and a blank value falls back to the default
  • Wiz inventory and Wiz compliance and STIG asset stubs record a virtual machine image as Software, as the default hardware type list intends
  • A Wiz vulnerabilities sync no longer opens issues titled with the CVE twice ("CVE-2016-9267: CVE-2016-9267") for findings Wiz names by their CVE alone: the title now reads the CVE once, followed by the start of the description or, without one, the affected package; the next sync also retitles existing issues of this kind and updates their Security Checks text to match
  • The AWS command help and integration guide now match what sync_findings, the Audit Manager and AWS Config syncs and the asset sync do
  • regscale aws sync_assets no longer adds an Athena workgroup asset with no account number when AWS STS refuses the call and the security plan holds that workgroup for more than one account
  • The child jobs of a split Wiz inventory sync now read Wiz at the same time, so a later child waits only for the first child's remaining read and the writes ahead of it instead of every other child's full read
  • regscale aws sync_assets no longer marks Inactive an asset that regscale aws sync_findings_and_assets created for a resource the inventory does not list
  • regscale aws sync_assets --account_id now skips Athena workgroups, with a warning, when the AWS credentials belong to a different account, instead of saving that account's workgroups under the scoped account
  • aws sync_findings and aws sync_findings_and_assets with --account_id or --tags now close no vulnerabilities, and a run without these options closes the ones no longer reported
  • regscale aws sync_config_compliance no longer fails its compliance lookup or evidence collection when AWS Config returns a rule with no name, and its warning shows such a rule by its ID or ARN
  • regscale aws sync_config_compliance no longer closes the vulnerabilities and issues of passing controls when AWS Config returns a rule with no ARN or no name. The run names the rule in a warning. When the run checks that rule, it also leaves mop-up off and does not update control assessments. Fix or delete the rule in AWS Config to close records again
  • When AWS refuses part of the S3 bucket listing, regscale aws sync_s3 now warns and does not cache the partial list
  • regscale aws inventory collect now warns about each resource type that AWS refused to list
  • If AWS refuses the GuardDuty detector listing or a detector's details, regscale aws sync_assets now keeps the GuardDuty detector assets Active instead of marking them Inactive
  • When regscale aws sync_assets cannot read a resource listing, it now keeps those assets Active and names the type in a warning
  • regscale aws sync_s3 keeps its earlier evidence file when AWS refuses to return one of a bucket's settings or a bucket cannot be read
  • regscale aws sync_s3 now replaces the evidence file it attached earlier for the same AWS account instead of adding one more every run, and keeps files others added, files in an approval and files from a narrowed or partial run
  • regscale gcp sync_assets with --asset-types or --labels no longer sets the plan's other GCP assets Inactive
  • regscale gcp collect_evidence (and regscale gcp sync_findings --generate-evidence) now replaces the evidence file on each evidence record instead of adding one more file every run, and keeps files a person or another account added, files in an approval and files in an earlier version of the record
  • The gcp sync_compliance, collect_evidence and sync_findings --generate-evidence commands no longer set the plan's other compliance assets Inactive
  • The gcp sync_compliance command now closes its own earlier Fail of a control once no active finding fails it, and keeps any other Fail
  • Fixed gcp sync_findings with --sources or --severity-filter closing the plan's GCP findings that the filter left out; only an unfiltered run closes findings that are no longer in SCC
  • Fixed GCP findings sync keeping a vulnerability linked to an asset its finding no longer names: when a finding moves to another asset, the sync now maps the vulnerability to the new asset and closes its link to the old one, unless the vulnerability is not Open, when the old link stays Open
  • gcp sync_compliance and gcp collect_evidence now fail with exit 6 (4 when GCP refuses the credentials, 1 when the findings cannot be read) instead of ending successfully with nothing assessed when Security Command Center cannot be reached or its findings cannot be read
  • regscale jira issues no longer changes the "Due Date:" line of an issue description to ISO form (2027-02-04T07:19:41Z) on a re-sync; the line keeps 2027-02-04 07:19:41, or 2027-02-04 for a Jira due date
  • jira tasks now applies Jira's title, description and due date to a task in the same run that closes it; before, a task edited and moved to Done in Jira closed with its old values until a later run fetched the issue again
  • regscale jira issues no longer changes an issue that a scanner integration wrote, so a closed AWS or GCP issue stays closed
  • regscale wiz compliance_report now replaces the report file on the project's evidence record instead of adding one more file every run, and keeps files a person or another account added, reports in an approval and reports in an earlier version of the record
  • regscale wiz compliance_report and the deprecated regscale wiz sync_compliance now close the issue of a control that the latest report passes
  • A Wiz vulnerabilities sync now updates an end-of-life finding's vulnerability in place when Wiz renames the finding, instead of adding a second vulnerability for the same finding
  • A repeat OCSF ingest no longer creates a second asset for a named resource that has no uid, and it updates only a same-named asset whose notes hold that resource
  • A repeat OCSF ingest keeps the asset status and other asset values a person set
  • Wiz commands that get a query answer with errors and no data now print one error line that names the query and the errors Wiz reported, instead of several tracebacks, and optional steps that skip on such an answer (asset enrichment, the representative-asset collapse) say what they skipped without a traceback
  • Vulnerabilities that an integration builds itself (the Qualys VM report and CIS imports) now carry a finding type like the rest: software vulnerability when there is a CVE, otherwise other.
  • SARIF, Qualys Total Cloud, OpenText WebInspect and DevSecOps SARIF and Wiz CLI imports, and Trivy or Grype imports whose files share no scan day, no longer send the CLI machine's local date as the scan history date when none was given, so RegScale dates the scan history by the import time
  • regscale gcp sync_compliance with CIS_GCP, PCI-DSS or SOC2 no longer fails controls from Security Command Center tags of another standard or version (CIS Controls, CIS benchmarks other than 2.0, PCI DSS 4.0, SOC 1)
  • The FedRAMP inventory workbook importer now stores a common "Asset Type" such as "Server" as "Physical Server" (the shared asset type aliases) instead of "Other", or as "Virtual Machine (VM)" when the row's Virtual column is Yes; values that match nothing are still stored as "Other" with a warning.
  • Microsoft Defender asset sync now marks public IP address resources as public-facing, as PS-929 specifies, instead of leaving the flag unset.
  • A Nessus or Tenable.sc plugin that lists several CVEs now takes its KEV-listed CVE (the one with the earliest KEV due date) as the vulnerability's CVE, so the issue's KEV list, the KEV due date and the VDR KEV clock still apply when the KEV CVE is not the plugin's first. A plugin with no KEV CVE keeps its first CVE, and the description still lists every CVE. The vulnerability is still keyed on the plugin ID, and the issue's identity stays on the plugin's first CVE when the primary CVE is swapped, so a CVE joining the KEV catalog does not reopen the issue.

Bug Fixes and Usability Enhancements

This release includes security hardening, patching, and fixes that improve reliability, performance, and overall application stability.

  • Version: 6.36.1.0 (minor release)
  • Date: 2026-11-16
  • Requires: RegScale 6.36.0.0 or newer

Summary

RegScale Orchestration Hub (ROH) automates the import and export of data between RegScale and other systems, using integrations and commands that you configure.

This release makes ROH refuse unencrypted connections in production. It also protects more of the API, the log stream and the webhook receiver. Classic integration jobs now run on a newer build of the improved, upgraded command line interface (CLI). ROH now keeps job names, descriptions and commands that use non-Latin characters, instead of turning them into question marks.

Several changes stop ROH from starting until you update its configuration. Read "Before you upgrade" first.

What's new

Encrypted connections in production

When ROH starts in production, it now tests its connections. If a test fails, ROH stops and the log names the setting to fix. The log never shows the setting's value. "Production" means your environment setting is production, which is the default, or that debug is off.

  • Redis. ROH refuses a remote Redis address that is not encrypted or has no password. It also refuses a Redis server certificate that does not match, including in cluster mode.
  • SQL Server. ROH requires an encrypted connection. It refuses a server certificate that does not match the host name. A certificate with a wildcard name, such as *.database.windows.net, now matches the server name.
  • Other RegScale services. ROH refuses an unencrypted web (http) address to another machine, for your ROH address and your RegScale address. A service account with an http RegScale address fails its runs until you correct the address.
  • Weak or sample keys. ROH will not start in production with a weak signing, encryption or metrics key, or with a placeholder key copied from an example file. These rules also apply when debug is on and the environment setting is not set.

Sign-in and sign-out

  • Signing out now ends your session on the server. Before, the Logout button cleared only the browser, and the session stayed valid until it expired.
  • Session tokens are stored encrypted. ROH keeps the RegScale session token in encrypted form in Redis. Each user with an open session signs in again once after the upgrade. Changing SECRETS_ENCRYPTION_KEY now ends all open sessions.
  • SAML sign-in is tied to the browser that starts it. Security Assertion Markup Language (SAML) sign-in now finishes only in the browser that began it. If someone finishes it in another browser, or reuses an earlier sign-in attempt, ROH answers "unauthorized" (401). SAML sign-in needs HTTPS (encrypted web traffic) because the cookie it uses is marked Secure.

Log stream and webhooks

  • The session token is no longer in the web address. A client of the log stream now sends the session token in its first message instead. ROH refuses a web address that carries a token. The access logs, including nginx's, no longer record anything after the question mark in a web address.
  • Webhook size and rate limits. A webhook request can be up to 1 MiB (1,048,576 bytes). One sender address can make 30 requests each second, with a burst of 60. ROH rejects a larger request (413) and any request above the rate (429).
  • Webhook checks now run before ROH reads the request body. An unknown webhook gets "not found" (404). A webhook that has a secret but no X-Webhook-Signature header gets "unauthorized" (401). Before, a body that was not valid JSON (a standard data format) got 422 first.
  • Old stored webhook headers are cleaned up. A database migration removes every stored webhook header except content-length, content-type, user-agent and x-request-id. Before, some rows written by an earlier rollout still held an Authorization header.

Approved hosts only

ROH now contacts only approved hosts. A source can use Amazon S3, Azure Blob Storage and Google APIs endpoints. A service account can use only the scheme, host and port of your RegScale address. An operator can approve more hosts. Saving a source or service account with any other host fails (400), and a job that uses a refused source or service account fails.

The source connection test now runs as a background task (a worker task). It stops after 55 seconds, and a tenant can run two tests at the same time.

Operator-only routes and a separate failed-task queue for each tenant

Seven administration routes now need an operator token. A tenant session no longer works for them. If no operator token is set, ROH answers "forbidden" (403) to every caller of these routes.

The dead letter queue is the list of tasks that failed for good. ROH now keeps a separate one for each tenant and app. A tenant sees and retries only its own entries. A new operator route, GET /api/v1/operator/dlq/tasks, lists entries across all tenants.

Classic integration jobs run on a newer build of the improved, upgraded CLI

Classic ROH integration jobs now run on a newer build of the improved, upgraded CLI.

  • Two new Tanium commands. tanium sync_benchmarks and tanium sync_stig_benchmarks sync Tanium benchmark findings and Defense Information Systems Agency Security Technical Implementation Guide (DISA STIG) findings to RegScale security checklists.
  • Tanium Sync CIS Benchmarks. Each checklist goes on the endpoint asset in the security plan that regscale_id names. If an endpoint has no asset in that plan, ROH skips its finding, even when another plan holds a copy of that asset. Run Sync Assets first.
  • Prisma software bill of materials (SBOM). When a job sets bulk_download to false, ROH now passes --no-bulk-download to the prisma sync_sbom command, and the command syncs one resource. It needs resource_id, resource_type (host or image) and asset_id (1 or more). Before, ROH sent nothing and the CLI downloaded in bulk.
  • More problems now show as failed jobs. The CLI now reports more errors through an exit code, which is the number a program returns when it finishes. ROH marks every exit code other than 0 as failed, so a job that passed before can now fail. For example, a missing required setting such as wizUrl ends with exit code 5, and a prisma sync_sbom download failure ends with exit code 4 or 6.
  • Microsoft Entra sync settings. The azure_ad sync commands now list five optional secrets: adTenantAdminGroups, adGroupOverrides, adIncludeDisabledUsers, adFollowNestedGroups and adCascadeRemoveThresholdPct. A secret set can now store them.

What's fixed

  • Job names, descriptions and commands keep non-Latin characters. Before, ROH saved a character outside the database code page, such as a Chinese character, as a question mark. Rows that ROH already changed stay changed. A job command can hold up to 512 characters, and an emoji counts as two.
  • Split jobs. A split job runs as several parallel pieces, called chunks. When it reached its limit on parallel chunks, it used to send a chunk with a negative limit. ROH now makes fewer, larger chunks instead.
  • Amazon Web Services (AWS) Elastic Block Store (EBS) volumes. An EBS volume asset now has the AWS account ID in its key, so Security Hub findings link to the asset.

Security hardening and patching

Minor bug fixes and enhancements

Before you upgrade

Read each row. A row marked "ROH will not start" stops the deployment until you act. The first three columns say what to do. The last column holds the settings and details for the person doing the work.

What changedWho is affectedWhat to doSettings and details
Four database changes (migrations) run at startup and cannot be reversed.Everyone who upgradesBack up the ROH database before you upgrade. After ROH starts, read the startup log for warnings.Migration 048 removes stored webhook headers that ROH does not keep. Migrations 049, 050 and 051 change the job command, description and name columns to Unicode types. If a migration cannot change a column, it writes a WARNING that starts with the migration number and changes nothing, and ROH still records it as done. Run the statement from the WARNING by hand. Migration 051 changes nothing when the change would give two jobs in one app the same name.
Containers must upgrade together.Everyone who upgradesUpgrade all ROH containers at the same time.During a rolling upgrade, a request that reaches a container of the earlier release can fail until all containers run 6.36.1.0.
A remote Redis needs encryption and a password. (ROH will not start otherwise.)Deployments with a remote RedisUse an encrypted Redis address with a password.Use a rediss:// address with a password in REDIS_URL, ROH_BROKER_URL and ROH_RESULT_BACKEND. Addresses for localhost, 127.0.0.1 and ::1 are exempt. If you cannot use encryption, set ROH_ALLOW_INSECURE_REDIS=true. A cluster that uses bare IP addresses for its nodes, such as Azure Managed Redis in open source software (OSS) cluster mode, needs REDIS_TLS_VERIFY=false. If your Redis certificate comes from a private certificate authority, set REDIS_TLS_CA_FILE to the CA bundle.
SQL Server needs encryption and a matching certificate. (ROH will not start otherwise.)Deployments that use SQL ServerUse a database host name that appears in the server certificate. An IP address does not match.The host name in DATABASE_URL must be a name in the server certificate. A certificate from a private CA fails with error 20002 and the log text "TDS server connection failed". To trust a private CA, mount the CA bundle and set SSL_CERT_FILE. This setting replaces the trust store for every OpenSSL client in the process, including HTTPS calls to RegScale. Or mount your own freetds.conf and set FREETDSCONF. Your file replaces the shipped file (/etc/roh/freetds.conf) and must set encryption = require. A deployment with its own Compose or orchestration files must mount the file itself. The user with ID 1001 must be able to read it. DATABASE_ALLOW_UNENCRYPTED=true turns off the SQL Server rules.
Service addresses must use https. (ROH will not start otherwise.)All deploymentsUse https addresses for your ROH, RegScale and Next Gen service addresses.Set https addresses for ROH_PUBLIC_URL, REGSCALE_URL and the two Next Gen service addresses, or set the matching opt-in setting, for example ROH_PUBLIC_URL_ALLOW_INSECURE_HTTP or REGSCALE_ALLOW_INSECURE_HTTP. Use an opt-in only when something else protects the connection, such as a service mesh.
Weak and sample keys are refused. (ROH will not start otherwise.)Docker Compose stacks that use the default keys from the example files, and any deployment with a weak keyReplace default and weak keys with strong ones before you start ROH.SECRET_KEY needs at least 32 characters. SECRETS_ENCRYPTION_KEY needs exactly 32. ROH_PREFIX_NAMESPACE must not be local_dev. SECRET_KEY, SECRETS_ENCRYPTION_KEY and METRICS_TOKEN must not hold a placeholder from an example file. An empty METRICS_TOKEN is accepted. These rules also apply when DEBUG is on and ENVIRONMENT is not set. If you change SECRET_KEY or SECRETS_ENCRYPTION_KEY, all sessions end, and ROH cannot decrypt credentials stored under the old key, so enter them again.
Seven routes need an operator token.OperatorsSet an operator token.Set ROH_OPERATOR_TOKEN to a random value of at least 32 characters. It must differ from METRICS_TOKEN, SECRET_KEY and SECRETS_ENCRYPTION_KEY. The routes are POST /api/v1/monitoring/recovery/run, GET /api/v1/monitoring/beat/schedule, POST /api/v1/monitoring/beat/trigger/{schedule_name}, GET /api/v1/monitoring/redis/info, POST /api/v1/commands/cache/clear, POST /api/v1/documentation/mapping/reload and POST /api/v1/analytics/metrics/cleanup. Callers send the token in the X-ROH-Operator-Token header. Entries that were in the dead letter queue before the upgrade no longer show in the tenant routes. Only the operator route with scope=legacy shows them.
Non-default hosts must be listed.Sources on MinIO, a private endpoint or a proxy, and service accounts for a RegScale host other than REGSCALE_URLList each extra host ROH should be allowed to contact, then restart the application and the workers.List a source's host in SOURCE_EGRESS_EXTRA_HOSTS. List a service account's RegScale host in REGSCALE_EXTRA_HOSTS. Each takes a comma-separated list, and a name such as *.example.com matches any host under example.com. At startup, ROH writes a WARNING with the IDs of stored sources it refuses. A source that resolves to a private address, such as an Azure Private Endpoint, needs its full host name in the list. A worker must read the celery queue for the connection test. Without one, every test ends after 60 seconds.
Webhook limits apply.Webhook sendersKeep each webhook request to 1 MiB or less, and no more than 30 requests each second from one address.Behind a load balancer, all senders share one address, so they share one limit.
The log stream and SAML sign-in changed.Custom clients only. The ROH web interface already sends the new messages.Update any custom log stream or SAML client.Send the session token in the first WebSocket message. The message is a JSON object with type auth and a token field. Send after_id without percent-encoding, and keep the first frame to 8,192 bytes or less. A larger frame closes the connection with code 1009, on every WebSocket route. nginx answers 403 for a log stream web address that has a token parameter or a percent sign in the query. A custom SAML client must keep the cookie that the init call sets and send it with the callback.
Classic jobs can now fail where they passed before.Everyone with classic jobsReview your classic jobs after the upgrade.A prisma sync_sbom job with bulk_download set to false needs resource_id, resource_type and asset_id.
Image builds check their inputs.Anyone who builds their own ROH imagePass a checksum with each pinned download version.The build fails if FRONTEND_VERSION is latest or a download does not match its digest. Pass FRONTEND_SHA256 with FRONTEND_VERSION, and the matching checksum setting with the version setting for the Next Gen tool download.

Summary

Trust Center 1.1.0 gives each person one role and one view, adds an Admin Console Home page, enhances FedRAMP vulnerability reporting, and enables you to serve Trust Center from your own domain.

Roles and RBAC

  • Trust Portal: a signed-in view for customers, partners and internal viewers, with a refreshed Request Access form and an admin approval queue.
  • Preview Mode: admins and operators now see each Portal as its audience does.
  • Simpler Navigation: portal settings move to Portals › Trust Portal, and the per-system Access tab is retired.

Admin Console Home

  • New landing page after sign-in presents cards summarizing access requests, document requests and expired evidence.
  • Portal tiles provide a weekly usage summary, and an "all caught up" state.

Note on the Beta Version

  • Features are subject to change in the releases leading up to general availability.

[7.1.0] - 2026-11-11

Added

  • Tanium sync_stig_benchmarks imports DISA STIG results as STIG checklists, and sync_benchmarks imports CIS benchmark and STIG results together, each under its own checklist tool
  • An overall file progress bar for the Nessus import, which also logs one line per file instead of three
  • Trivy and Grype imports now remember which scan files were imported and skip them on the next run, with --reprocess to import everything again; a run that skips scan files does not close findings
  • Burp CSV and XLSX import command, regscale burp import_burp_csv, which imports a Burp scan held as a CSV or an XLSX (the Burp XML export with one row for each issue) and offers a --dry-run preview
  • Settings tab in the interactive GUI for browsing and editing every CLI configuration key, grouped by integration with secrets masked
  • New regscale risks export_risks and import_risks commands copy risks, with their attachments, between RegScale instances. Each risk's risk model is matched by title (or set with --risk_model_id), each imported risk is owned by the importing user (or the user given with --risk_owner_id), and re-running an import does not duplicate risks.
  • Databricks sync_assets and sync_findings commands that load the CIS checks and CVE findings of a Databricks export into a security plan as assets, vulnerabilities and issues, either into one plan or routed to many plans by an export column and a customer-supplied CSV; a CVE finding that the export marks closed is written closed, its assets carry the operating system and internet exposure the export gives, and --severity-field picks which export column sets each finding's severity
  • Option to have tenable io sync_findings read DISA STIG audit results from the Tenable.io compliance export and send them to the matching control objectives, off by default (tenableIoStigComplianceExport)
  • Vulnerabilities synced by scanner integrations now carry a finding type so non-CVE findings can be told apart from CVEs
  • New wizHeldFindingsQuietSeconds setting (default 30 seconds) sets how long wiz vulnerabilities waits, once no new asset appears, for a concurrent wiz inventory to write the assets of findings it holds

Changed

  • Tanium sync_cis_benchmarks now writes checklists only to assets in the security plan given by --regscale_id or in one of that plan's components; endpoints whose assets are all in other plans or their components are skipped instead of being written to a copy elsewhere, and the run reports how many were skipped
  • A removed init.yaml key the CLI no longer reads (the gRPC ingest keys, cisaKev, jiraDefaultIssueType, jiraDefaultPriority) now logs a warning on every run that names what to use instead, rather than being ignored silently, unless it still holds the value an older CLI wrote into every init.yaml; regscale config validate reports these keys as removed_key
  • Trust Center requests are now refused if they would leave the configured Trust Center host, API errors no longer echo the request's query string, and each request writes one debug log line without headers, bodies or query strings
  • Memory use when importing a large SARIF file for compliance mapping is roughly halved, because the file is now read in a stream instead of being loaded whole
  • Breaking change for custom scanner integrations: the ScannerIntegration flag that turns on closing outdated findings is now named close_outdated_findings_enabled, and the old close_outdated_findings name is no longer accepted as a flag because it is the inherited close method. Passing it to set_keys or to sync_findings/sync_assets on a ScannerIntegration subclass now raises TypeError. Two cases fail silently instead, so audit custom code for them: a class-level or instance-level close_outdated_findings = True assignment overwrites the inherited close method with a bool, and reading close_outdated_findings as a flag (if x.close_outdated_findings:) now returns the method, which is always truthy. Use close_outdated_findings_enabled for both.
  • A refused token now names its likely cause: an expired token says when it expired, and an unexpired one lists the RegScale 6.36 events that revoke tokens (password change or reset, administrator sign-out or deactivation, local sign-in turned off) with the fix for a service account or a person. A refused XCCDF benchmark import now logs the server's status, error code and message, such as the 409 that asks you to delete a benchmark with assessment results before replacing it
  • Nessus import reads large scan files faster and no longer loads the CPE dictionary when a scan names no CPE
  • Scanner imports make fewer requests to RegScale at start-up and spend less time per finding
  • The Nessus import is faster on folders with many files because it frees memory after a set amount of input instead of after every file
  • Trivy and Grype findings no longer start with the CVE in their title, and carry more of the scan data where the scan provides them: references, layers, CWE ids, EPSS and known-exploited flags, fix advice and image details
  • regscale jobs names each run's log file with its start time first, in UTC (20261108T070012Z-<job>-<run id>.log), so a plain ls of job_logs/ lists runs in the order they started. Logs from earlier runs keep their old names
  • Databricks sync_findings --mop-up now closes CIS checks and CVE findings separately, each under its own source (Databricks CIS Benchmark and Databricks CVE), so an export of one type no longer closes the other type's records, sync_assets --mop-up is refused unless the export holds both types, because the two types share hosts, and a plan that already holds issues under the earlier source Databricks gets a second set of issues on its first run, so close or delete the old ones first
  • Synqly connector prompts hide any value that the Synqly schema marks as a password, such as a client secret
  • Aqua imports use less memory on large reports
  • Each Xray vulnerability is now stored under a key that includes its CVE. The first Xray import after the upgrade creates new vulnerabilities with a new first seen date and closes the old ones, which keep their history.
  • On defender export_resources, -n now means --query_name only, so a script that used -n to skip the upload must use --no_upload instead
  • When a scanner sync stops because its vendor source fails partway through, the findings read before the failure are now sent without closing any findings, and a later error while sending them is logged at ERROR so the vendor failure still sets the exit code
  • If all rules of a control are in the conformance pack, aws sync_config_compliance --conformance_pack_name now closes its fixed findings and issues

Fixed

  • The interactive terminal GUI no longer fills the console with random characters when a file upload fails and the server returns a non-text error response
  • jcamValueMapping accepts @drop for a value, declining it: the value writes nothing and leaves the destination as it is, instead of being written literally as the text @drop; a value mapped to @unmapped still works and is noted once per run in favour of @drop
  • jcam import_ssp no longer stops for every plan when JCAM returns an undated SORN notice beside dated ones; undated notices go in the last slots, and an unexpected error in one plan's privacy data now fails only that plan
  • jcam import_ssp now sets each child plan's parent from JCAM's FISMA rollup, including a first parent and a change of parent, instead of every change being refused, and leaves the parent alone when JCAM's parent system has no RegScale plan instead of writing 0
  • jcam import_ssp no longer creates a plan a second time, without its compliance setting, in the front-matter step after RegScale refused to create it; the system is skipped and the refusal stays recorded
  • regscale jcam import_parameters counts a failed read of a plan's controls, or of a control's parameters, as a failure in the run summary, instead of treating the plan as having no controls or the control as having no parameters
  • jcam import_assessments and import_ssp count each assessment or system role RegScale's batch create did not return as a failure, instead of counting a refused batch or item as written
  • tenable nessus import_nessus now sets each asset's type from the host instead of creating every asset as "Other": an asset-type host tag (any asset type value, case and spacing ignored) wins, otherwise the standard system-type property is mapped (router, switch, firewall, load-balancer, hypervisor, mobile, general-purpose, and now embedded and printer as Appliance). The asset category follows the type (Software for application, database, container and cloud types, Hardware otherwise), assets created for findings use the same type, and a host with no type information no longer resets the type already stored for it when a newer scan is imported (its category is still sent as Hardware) (requires RegScale platform 6.35.1.0 or newer)
  • jcam import_poam matches an existing POA&M only among the importing plan's issues, instead of taking any issue in the tenant with the same otherIdentifier and moving it onto the plan
  • Tanium Cloud CIS benchmark checklists now show the rule title and the last scan date instead of only the rule ID
  • ServiceNow incident sync now copies attachments between incidents and RegScale issues instead of always reporting zero
  • AWS asset console links now open the AWS GovCloud console for resources in GovCloud regions
  • AWS Lambda layer version assets are now named by their layer and version instead of the bare version number
  • Jira evidence sync lists the Jira Tasks named in a repeated-failure log line in sorted order, so the line reads the same on every run
  • axonius_v2 commands give up on an Axonius request that times out after one attempt instead of four, and exit 6 with one timestamped log line naming the endpoint, the limit and axoniusTimeout, instead of a traceback or an unlogged error; the connect wait is capped at 30 seconds, as for other vendors, and axoniusTimeout is the read limit
  • jcam import_ssp skips JCAM's placeholder information type 'Other: Other' unless the tenant has a type of that name, instead of counting a plan that lists only the placeholder as failed and offering it in missing_info_types.json
  • The final step of a long import, where the RegScale server finishes the sync run, now logs when it starts, every 30 seconds while it runs and when it is done, instead of going silent for minutes
  • Interactive imports that send many batches, such as a Nessus folder import, no longer fill the terminal with finished sync progress bars that pushed the log output off screen part-way through the run
  • Trivy and Grype findings are dated from their scan file instead of the import time, and with import_scans first seen and last seen no longer move the wrong way when a newer or older scan is imported
  • --s3-bucket and --s3-prefix now read the bucket for Trivy and Grype imports, using the AWS profile or keys given on the command line
  • Compliance integrations, including AWS Security Hub compliance and Wiz compliance_report, count a compliance item that fails to process as a partial failure, so the run exits 3 under REGSCALE_STRICT_EXIT_CODES=1 instead of 0
  • Commands run without a terminal (regscale jobs run-due, cron, a redirected shell) now log a progress line at INFO at most once a minute during long loops, and a closing line with the count and elapsed time, where they used to stay silent until the loop ended
  • regscale jcam import_controls creates a plan's new control implementations 100 per request instead of one request each, and sends a control the report repeats once
  • Databricks sync_findings and sync_assets with --regscale-ssp-id use less memory on a large export, because they no longer keep one object for every distinct host of the whole file while they stream it
  • Tenable.io sync_compliance_controls no longer stops on a compliance check with a null see_also, description or other descriptive field, and skips a check it cannot read instead of ending the run
  • FedRAMP Appendix A import no longer stops when a control in the document has no implementation status selected
  • The FedRAMP POA&M Excel import now accepts workbooks built from template v3.x. It no longer requires the removed Resources Required, Planned Milestones, Milestone Changes and Operational Requirement columns, and it reads Remediation Plan in place of Overall Remediation Plan. It skips a POA&M sheet that has no POAM ID header, with a warning, instead of stopping. Typing a quoted answer such as 'skip' at the mapping prompt no longer loops.
  • Trivy, Grype, SARIF, OpenText, Qualys Total Cloud and Tenable SC file imports no longer fail with "cannot pickle '_queue.SimpleQueue' object" when the staged findings file is larger than 500 MB
  • Compliance syncs now fill in the control and security plan on each control assessment, and the control and assessment on issues whose finding maps to a single control. A new issue gets its assessment on the first sync and its control on the next
  • Critical RegScale issues synced to Salesforce now create cases with High priority instead of Low
  • JCAM milestone import reports an unrecognized or undated completed status instead of guessing, and export sends a milestone's own status to JCAM when it is one JCAM uses instead of always Completed or Not Started
  • JCAM milestone import no longer gives a milestone JCAM sent no due date the import time; RegScale sets the date from the POA&M's severity
  • Reminder emails no longer fail for users who have no upcoming or overdue items, and the run now reports when nobody has any
  • Qualys scanner no longer fails on a finding with no detection data and now gives such findings a unique id
  • regscale tenable io sync_compliance_controls now creates a control implementation for each matched control that passed or failed a check and that the security plan does not have, gives an enhancement control (for example AC-2(1)) the status Tenable.io reported, and exits 1 when it matches controls but RegScale saves none of them
  • The stored size of an uploaded file, such as a Jira attachment or evidence file, is now its byte count
  • AWS sync_assets on plans with assets in more than one region now marks a deleted resource as not in the latest scan and keeps edits made to it during the run, and the AWS compliance commands keep an Inactive status a user set and give S3 buckets the Cloud Storage type without replacing their description
  • On platforms before 6.36.0.0, regscale aws sync_config_compliance now closes the issues that CLI 6.x stored under the bare AWS account ID when it retires those records, instead of leaving them open, with issueCreation: Consolidated (the default)
  • regscale aws sync_config_compliance --conformance_pack_name no longer closes the findings and issues of controls that pass, so the findings and issues of failing rules outside the conformance pack stay open
  • Scan History Date Scanned now shows the scan date instead of the import time for integrations that import vulnerabilities in batches
  • Asset syncs no longer log a duplicate asset that RegScale saved from a later row of the same batch as "not saved", and the "Successfully submitted" and "asset(s) synced" counts now show distinct assets
  • Synqly connector syncs create their Synqly account and integration with the installed Synqly SDK
  • Amazon Inspector syncs through Synqly build an AWS access key credential and ask for the access key ID and the secret access key, instead of stopping with an attribute error
  • Synqly ticketing syncs create tickets and upload attachments with the installed Synqly SDK
  • The error that follows a rejected Synqly integration setup no longer shows the credential values that Synqly echoes back
  • Synqly vulnerability syncs such as Amazon Inspector, Wiz and Tenable Security Center no longer stop with a filter validation error before they fetch findings
  • The file comparison command now names the supported file types, CSV, XLSX or XLS, when it refuses a pair of files
  • If another import saves the same plan, an asset import no longer adds a second copy of an asset. It reports the asset as not saved, with the reason
  • The FedRAMP OSCAL SSP import now ends with one clear error line and exit code 6, or 4 when the host refuses access, when the host of the profile that the SSP names cannot be reached or answers with an HTTP error, instead of a traceback
  • The FedRAMP OSCAL SSP import now reads a profile that its host sends compressed, such as the FedRAMP profiles published on GitHub, instead of failing with a decoding error
  • The FedRAMP OSCAL SSP import now ends with one clear error line and exit code 1, instead of a traceback, when the profile that the SSP names is not a readable JSON document, and it no longer leaves a copy of the profile in the system temporary directory
  • The eMASS POA&M workbook import now stores each POA&M with an Issue status the tenant offers (Ongoing as Open, Completed as Closed with its completion date, Risk Accepted as Exception/Waiver or Open, and Open or Closed as Ongoing or Completed on a tenant that uses the eMASS statuses) instead of losing rows.
  • Tenable.io commands now honor the sslVerify setting, and Tenable.io and Tenable.sc commands now verify against the CA bundle set in customCaCert
  • Tenable.io asset sync stores an asset's MAC addresses in the same order on every run
  • tenable io info shows the Tenable.io user, account email and scanners instead of always reporting a connection error, and exits with a non-zero code when it cannot get them
  • regscale tenable sc list_tags now lists the Tenable.sc tags. It asked pyTenable for a tags endpoint that does not exist, logged an error and exited 0. It now reads GET /rest/tag, prints "No tags found in Tenable.sc." when there are none, and exits 4 when Tenable.sc refuses the keys, 6 when it does not answer or reports an error, and 1 when the reply is not a list of tags.
  • Prisma CSV import removes repeated rows before importing, so the logged counts are of distinct findings
  • Wiz add_report_evidence now fails with an error naming the refused upload, instead of logging that the file was uploaded
  • Jira commands that use basic authentication now honor the sslVerify setting, and a blank sslVerify key verifies the certificate
  • OpenSCAP imports no longer leave two open issues for each failed rule, and the one issue that remains is linked to the vulnerability but not to the affected controls
  • axonius_v2 commands print one error line that names the Axonius host and the failure, and exit with code 1, instead of a traceback when Axonius cannot be reached, drops the connection or answers with an error or an unreadable reply; a timeout still exits with code 6
  • axonius_v2 now prints one error line that names the input, and writes nothing, when a saved query config file is missing or invalid, a saved query is not in Axonius, a regscaleSspId is not an integer, or --offset or --limit is negative
  • Axonius asset sync no longer imports nothing when one device has no preferred hostname or name, or one flat-file row has no CDM Unique ID; such a device is synced under another identifier, and a device with no identifier at all is skipped with a warning that names it
  • Fixed axonius sync_assets --regscale_ssp_id to stop with an error when the security plan does not exist instead of writing assets under it
  • Axonius asset sync now stops with an error and a non-zero exit code when RegScale refuses an asset batch, for example when the token has expired, instead of reporting completion
  • Axonius pull_data_axonius gives up when Axonius sends nothing for the axoniusTimeout setting (exit 6) instead of a fixed 3600 seconds, and reports a refused connection, an unknown host or a reply that is not a JSON object as one error line naming the host (exit 1) instead of a traceback
  • Axonius pull_data downloads only the S3 object that the key names and fails when no object, or several objects, match the key
  • axonius combine_data_parts merges a part that holds one JSON object as one record instead of writing its field names, and stops with a usage error for a part that is not a list or object
  • axonius file commands and combine_data_parts report a missing, unreadable, non-UTF-8, empty, non-JSON or non-list input file, and a tenant without the "FISMA Id" security plan field, as one Error: line with exit code 1 instead of a traceback; the sync commands now check the input file before they query RegScale
  • Report a missing security plan in the Axonius V2 sync commands as one error line instead of a traceback
  • prisma sync_sbom (and panw sync_sbom) now fails when the SBOM download fails or the archive cannot be read, with exit 4 when Prisma Cloud refuses the credentials and exit 6 otherwise, instead of reporting success and exiting 0
  • The Prisma Cloud and Palo Alto sync_sbom commands now accept --no-bulk-download to sync the SBOM of one host or image
  • Prisma authenticate (and panw authenticate) now stops with exit status 4 when the login reply holds no usable token and with exit status 6 when the reply is not JSON, caches nothing, and ignores a tokenless session that an older release cached
  • Prisma Cloud syncs now wait for each Retry-After value, at most 30 seconds, log every retry wait and fail at once when the console asks for a longer wait
  • Burp Suite imports no longer stop when an issue has no background or a request and response pair has no request
  • Importing a Burp Suite export now maps a vulnerability to each host that reports an issue type, where only the first host got it; the issue names every such host, and the stored vulnerability key does not change
  • The Burp import now names a file that is empty, cut short or not a Burp export, and exits with an error when no file in the folder could be imported
  • Burp import no longer prints a time zone warning for an export time that ends in a zone abbreviation such as EDT
  • The Acunetix and Prisma CSV imports, and the other flat-file imports, now print one line for each file they cannot read and exit 1 when no file in the folder could be imported
  • Importing a JFrog Xray report now stores one vulnerability for each CVE that a finding lists, where it stored only the first
  • Fix the JFrog Xray import storing a finding as Low when its CVE has no CVSS v3 score: it now uses the severity Xray gives the finding, or the band of the CVSS v2 score
  • Fix the JFrog Xray import skipping a finding that has no CVE identifier. The import now stores such a finding under its Xray issue id. A warning now names each finding that the import skips
  • Importing a JFrog Xray or AWS Inspector JSON report with no findings now completes and says the report holds no findings instead of asking for a column mapping or failing, and an Xray report page that lists none of its rows is skipped so it closes no findings
  • JFrog Xray import of a report that holds one finding as a single object instead of a list
  • The Xray, Snyk and other JSON imports now name a .json file that is empty, cut short, not valid JSON, not readable or not a list of records, skip it, and import the other reports in the folder. A JSON report that starts with a UTF-8 byte-order mark is now read
  • Prisma CSV import keeps findings identified by a Prisma Cloud or GitHub advisory ID instead of skipping them
  • Prisma CSV import stores a finding with no CVSS score, or a score of 0.0, at the severity in its Severity column instead of Low
  • A mapping file that is empty, not UTF-8 text, not valid JSON or not a JSON object now stops show_mapping with a message that names the file, instead of a traceback or a printed value
  • Scanner file imports and other commands that check their security plan or component now stop with an authentication error when RegScale refuses the account (exit status 4, or 78 when the token cannot be renewed), instead of reporting an invalid plan and exiting 0
  • Defender commands now report that Azure refused the credentials, with Azure's error description, instead of a missing access_token key and a traceback; the exit code stays 1
  • Defender export_resources now reports the HTTP status and the request that failed when Azure refuses or fails the saved query request, instead of a traceback
  • defender export_resources -n "<name>" now exports the named saved query and no longer prints five duplicate-option warnings
  • Microsoft Defender alerts are no longer skipped when an activity time cannot be reformatted, and a time with an offset now shows its UTC date
  • Entra evidence collection logs one error line for each evidence type that Microsoft Graph refuses, in place of a traceback and a bare exit code
  • regscale emass_api authenticate, push_poam and sync_system now stop after one wait when eMASS does not answer, print the line eMASS at <host> did not answer in time (<n> seconds). and exit 6. push_poam no longer sends a new POA&M when its search for an existing POA&M gets no answer, a bidirectional sync_poams prints how many POA&Ms it pushed before the timeout, and authenticate no longer exits 0 when its connection test fails
  • regscale emass_api commands no longer wait without a limit when eMASS accepts the connection and does not answer. The CLI now uses the emass.timeout setting (or EMASS_API_TIMEOUT, default 300 seconds) for each eMASS request. After that time the command prints one line, eMASS at <host> did not answer in time (<n> seconds)., and exits 6. When the request was a change, the line adds eMASS may have applied the change., so check eMASS before you run the command again. A command that sends many items, such as sync_controls or upload_artifacts_bulk, stops at the first timeout. The time limit applies to each phase of a request, not to the whole request, so a server that sends a little data at a time can still keep a command waiting. A large artifact upload to a slow eMASS server can now stop at 300 seconds; raise emass.timeout for it. A value of emass.timeout that is not a positive number, such as 0, -5 or abc, uses 300 seconds and logs a warning.
  • regscale emass_api sync_controls no longer exits 0 when no control synced because eMASS could not be reached or refused the request. When no control synced and at least one control failed because of eMASS, the command prints one error line with the eMASS host and exits 4 if eMASS answered 401 Unauthorized or 403 Forbidden for any control, and 6 for an unreachable host, a TLS failure, any other error status or an answer it cannot read. It used to print [OK] Control sync complete with every control counted as an error. When eMASS refuses only the content of controls (400 Bad Request, 409 Conflict, 422 Unprocessable Content), or when some controls synced, the failed controls stay partial failures: the command prints how many controls were not synced, with the synced and error counts, and no longer prints [OK]. The DEBUG log message for a failed control push no longer carries the eMASS response body.
  • regscale emass_api upload_artifacts_bulk and add_milestones_bulk no longer print tracebacks or the eMASS response headers and body when eMASS fails, sync_poams prints one error line that names the eMASS host when its pull fails, and import_xml prints one error line without a traceback for a file that is not XML
  • The regscale emass_api commands now print one error line that names the problem and the eMASS host, for example eMASS at emass.example.mil rejected the API key: 401 Unauthorized., instead of one to three tracebacks. emass_api test_connection no longer ends with an empty Error: line and Aborted!. The line and the warnings never carry the API key, the response headers or the response body. A 401 Unauthorized or 403 Forbidden answer exits 4. An unreachable host, a failed TLS connection, any other HTTP error status or an unreadable answer exits 6. Both used to exit 1. A local or RegScale error still exits 1.
  • The regscale emass_api commands delete_milestones, add_milestones_bulk and upload_artifacts_bulk accept --yes (or -y) to skip the confirmation question, so a scheduler, a pipeline or the Orchestration Hub can run them. Without --yes and without a terminal, each command now prints one Error: line that names --yes and exits 2 before it calls eMASS, instead of a traceback and Aborted!.
  • regscale emass import_slcm now reports a file that is empty or is not an Excel workbook as one error line that names the file, and exits 1, instead of printing a traceback
  • The eMASS import commands and detect_workbook now report a file that they cannot read as one error line that names the file and exit 1, instead of a traceback or exit 0
  • regscale emass import_slcm now exits 1 when the security plan or the catalogue is not found, 4 when RegScale refuses the lookup and 6 when RegScale does not answer it, instead of exiting 0 after logging "Aborting!"
  • If a control has the Implementation Status Manually Inherited, a blank status or an unknown status, the regscale emass import_slcm command no longer marks it Not Applicable. Manually Inherited is now saved as Manually Inherited, as regscale emass import_slcm_workbook does. The Implementation Status is matched without regard to case. A blank or unknown status keeps the control status and its N/A Justification, and logs a warning that names the control. The command applies Not Applicable from the Implementation Status or Compliance Status column only with an N/A Justification. Without one, the control keeps its status and the command logs a warning.
  • The eMASS hardware and software inventory import now reports each row it skips and counts each asset RegScale refuses as an error
  • eMASS POA&M workbook import now stops and names the rows that share a POA&M Item ID, instead of merging them into one issue
  • eMASS control test results import now names the row and value of an unreadable Date Tested when the CCI does not resolve
  • eMASS control test results import no longer sends an unreadable Date Tested to RegScale and names the CCI number in its errors
  • OpenSCAP import_csv now names a CSV file that is cut off, has a header that matches no supported layout, is not UTF-8 or has a field over the CSV size limit, and exits non-zero instead of importing part of the file or nothing
  • OpenSCAP imports now name each missing, empty, truncated or non-XML scan file, and a folder with no scan file, and exit non-zero, while the other files in the folder are still imported
  • OpenSCAP CSV import now reads files that start with a UTF-8 byte-order mark, such as files saved by Excel as CSV UTF-8, and imports the same findings as the file without the mark
  • cci_importer prints one error line that names the file, with no traceback, when the CCI file is empty, truncated, not XML or cannot be read
  • regscale stig cci_assessment and regscale stig_mapper run now stop with an error and a non-zero exit status when the security plan does not exist, in place of reporting that they found nothing to do
  • The jira_dc commands now log one error line that names the problem and the Jira Data Center URL, with exit code 4, 5 or 6, when the first request to Jira Data Center fails, and a server that does not answer fails after jiraDcTimeoutSeconds instead of about six minutes
  • Jira commands jira issues and jira tasks now end with one error line and a non-zero exit code when Jira refuses the credentials, cannot be reached or answers with an error, and they report an unset jiraUrl, instead of printing a traceback
  • Qualys scan import no longer stops when a row's severity is empty, not a number or outside 1 to 5; it logs a warning naming the row and imports that row at low severity
  • The example Qualys URL in the init.yaml template is now a host only, and an existing init.yaml that keeps the old example path no longer doubles it or triggers config validation warnings
  • Qualys diagnostics and CSV validation now exit with a non-zero status when a check fails, and a failed request no longer prints a traceback
  • The Qualys import_policy command now reports a one-line error and exits with status 2 when the file is not a Qualys policy export or an option value is not valid, and no longer creates an Untitled Policy from a JSON object that has no policy ID
  • Qualys import-vm-report now stops with a usage error when both --report-file and --report-title are given, instead of silently importing the file and ignoring the title
  • A mapping file that is not valid JSON, not UTF-8 text or of the wrong shape now stops the import with exit status 2 and one line that names the file, instead of a traceback
  • Import a Qualys VM scan row that has no QG Host ID as an asset keyed on its IP address, and skip a row with neither
  • A Wiz token URL that ends in any mix of slashes and whitespace, such as a space or a tab followed by a slash, is now accepted, as one that ends in a slash or a space already was
  • Wiz commands now name the token endpoint when its answer is not a token, instead of a redacted placeholder
  • regscale wiz attach_sbom now attaches an SBOM whose document has no version, and when RegScale refuses one SBOM it still attaches the others and then exits 1
  • A compliance sync that stops on a usage error, such as a Wiz compliance report that is not UTF-8 text, prints one "Error:" line and no longer also logs the same message with a traceback
  • Wiz inventory, issues and vulnerabilities help now names the --filter_by_override option and shows an example that is valid JSON
  • The deprecated wiz sync_compliance command no longer lists --list-frameworks in its help, and using that option now stops with a usage error (exit code 2) that points to wiz compliance_report instead of exiting 0 after printing a message
  • The tenable was sync_findings command no longer exits 1 before it sends a request to Tenable.io when the security plan already holds assets
  • A Tenable.io --tags value without key:value pairs, an --asset-filter-file that is not a JSON object, and a Tenable.sc sync_jsonl --mapping_file that is not a JSON object now end with a usage error that names the option and the expected format. A file passed as a pipe, such as a shell process substitution, still works
  • Tenable.io CIS checklist syncs and Tenable.io STIG compliance results now send asset tag filters in the form Tenable.io accepts, so a run with --tags no longer fails and syncs nothing
  • Tenable.sc file import with --file-path now stops with an error and a non-zero exit status when no file in the folder can be read as Tenable.sc data
  • Negative --offset and --limit values are now refused with a usage error instead of stopping a sync or processing the wrong records
  • Log redaction no longer hides the start of a connection error after a sign-in URL that ends in /token, so "[Errno 111] Connection refused" stays readable
  • When Microsoft Defender keeps failing or does not answer, the error now names the request URL again, without its query string or any credential
  • A request that keeps failing with an HTTP error such as 500 Internal Server Error or 429 Too Many Requests now logs one error without a Python traceback, and Microsoft Defender reports a sustained failure once
  • A missing required setting, such as wizUrl, now ends a command with one error line and exit code 5 instead of a Python traceback
  • tenable io sync_compliance_controls keeps the control owner of each control implementation it updates
  • Qualys import-vm-report now saves Re-Opened and Fixed detections with a RegScale vulnerability status, and exits non-zero when it saves no asset or vulnerability or a request fails
  • The eMASS asset, POA&M and test result integration classes now read the eMASS responses and sync the systems, POA&Ms and test results they return instead of failing with an error
  • Tanium benchmark and ports syncs stop with an error when RegScale refuses or fails the read of the plan's components, instead of skipping the assets under those components
  • Tanium ports and protocols sync in an Orchestration Hub split job now reads ports for the same endpoints that the job's asset step writes
  • Tanium ports and protocols sync now writes ports only to the endpoint's asset in the selected plan or its components, not to copies of the endpoint in other or deleted plans
  • eMASS and FedRAMP imports that update control implementations no longer reset the inherited record permissions or clear the last assessment, the parent control and the approval state of those control implementations
  • If two rows of an asset sync differ only in letter case, the sync writes the software and SBOM of their asset once
  • The jira_dc sync_tasks command no longer creates two RegScale tasks when the Jira search returns the same ticket on more than one result page
  • FedRAMP OSCAL SSP imports now store the ports and protocols of each component, which RegScale rejected before because their Service and Used By values were empty
  • Compliance syncs on RegScale 6.36+ read each control's newest same-day assessment from POST /api/v3/assessments/byParents, so a control assessed more than once today merges with its latest result instead of its earliest.
  • AWS sync_assets now shows the valid from date, the valid until date and the number of resources in use on each ACM certificate asset
  • Qualys sync keeps a consolidated issue and vulnerability open while any host or web application still reports the vulnerability as open
  • Qualys WAS findings that Qualys reports as fixed now close their issue and vulnerability during sync
  • env_info now prints its full report in the container image, under an unknown user id and in a read-only directory
  • Pipeline report_scan no longer stops before it creates the issues for a failed scan when the SSP-level control update fails, and RegScale no longer refuses that update because the control origination is missing
  • Pipeline setup_component and report_scan now link the repository component to the security plan, so it appears in the plan's Components tab
  • regscale wiz vulnerabilities now names every finding it did not write when it exits 1 because regscale wiz inventory had not written the plan, one line for each finding with its asset
  • aws sync_findings without --generate-evidence or --format now sends the findings it read before a failed Security Hub page, instead of none of them
  • A scanner sync that the server did not start, because another sync of the same source was still running, no longer sends its checklist results or changes the status of a control or control objective
  • Checklist roll-ups send only valid assessed dates, and a Tenable STIG result whose scan date is missing or not a valid date is dated by its last seen date, so one bad date no longer stops every result in the sync from updating control status
  • Tenable STIG imports send a passing check's vulnerability as closed and any other result as open, instead of sending the check result, which the server rejected with a warning for every passing check
  • Wiz vulnerabilities and issues incremental syncs run with --offset or --limit no longer move the saved pull watermark past findings they did not write
  • Stop logging an error and stack trace when the catalog update cannot read the platform feature flags
  • A Tenable.io asset or vulnerability sync now stops before it sends any record from a cached export that is corrupt, and an export that stops partway is no longer cached for the next run
  • A truncated Tenable.io asset or vulnerability export cache is now removed and fetched again on the next run instead of failing every run until it expires
  • Updated the pyasn1 dependency to a release that fixes known security vulnerabilities
  • BigQuery asset sync now reports a rejected query as a single error line instead of a Python traceback
  • Issues from a component compliance report are now saved under the component and no longer rejected as a missing security plan
  • A Qualys CSV import whose column changes type after the first 10,000 rows now stops with a one-line message that names the file instead of a traceback
  • A Qualys file import with a negative --skip_rows value, or one that reads a title row as the column headers, now stops with a message that names --skip_rows instead of a traceback
  • Tenable compliance sync gives a failing control its planned implementation date and steps to implement, and sets a passing control to Fully Implemented, on platforms that take bulk control updates
  • Qualys import_total_cloud no longer logs that it imported findings RegScale refused, and reports how many vulnerabilities RegScale stored
  • Re-importing a Qualys VMDR report into a security plan now updates each host's existing asset instead of creating a second asset for every host
  • The eMASS POA&M workbook import now stores the controls, devices affected, impact description and security plan of each POA&M it creates, and fills the false positive, operational requirement, auto-approved and risk adjustment flags with No when they are empty. A row the batch upload refuses, such as a Risk Accepted row stored as Exception/Waiver, is now written on its own, or reported with the server's reason, instead of being dropped with "see the log".
  • FedRAMP CIS/CRM imports store every control and part inheritance model when the RegScale platform does not accept them all in one bulk request, and fail with an error line when RegScale refuses a model
  • Evidence and scan results now map to every control of a security plan with more than 500 control implementations when RegScale leaves out the next page marker on a full page
  • Evidence and scan results now map to control enhancements on a RegScale platform with the bulk compliance routes, and evidence also maps when the source names a control enhancement in the AC-2(1) form
  • On RegScale 6.36 and later, a compliance sync no longer replaces the result of an assessment from another day
  • Control implementation syncs and imports on RegScale 6.36.0.0 or later now retry or report a control implementation that the bulk update did not write, instead of finishing as if it succeeded
  • Wiz syncs that run longer than the Wiz access token lasts get a new token and keep going, instead of failing with an HTTP 401 and exit code 4

[7.1.0] - 2026-11-11

Added

  • Tanium sync_stig_benchmarks imports DISA STIG results as STIG checklists, and sync_benchmarks imports CIS benchmark and STIG results together, each under its own checklist tool
  • An overall file progress bar for the Nessus import, which also logs one line per file instead of three
  • Trivy and Grype imports now remember which scan files were imported and skip them on the next run, with --reprocess to import everything again; a run that skips scan files does not close findings
  • Burp CSV and XLSX import command, regscale burp import_burp_csv, which imports a Burp scan held as a CSV or an XLSX (the Burp XML export with one row for each issue) and offers a --dry-run preview
  • Settings tab in the interactive GUI for browsing and editing every CLI configuration key, grouped by integration with secrets masked
  • New regscale risks export_risks and import_risks commands copy risks, with their attachments, between RegScale instances. Each risk's risk model is matched by title (or set with --risk_model_id), each imported risk is owned by the importing user (or the user given with --risk_owner_id), and re-running an import does not duplicate risks.
  • Databricks sync_assets and sync_findings commands that load the CIS checks and CVE findings of a Databricks export into a security plan as assets, vulnerabilities and issues, either into one plan or routed to many plans by an export column and a customer-supplied CSV; a CVE finding that the export marks closed is written closed, its assets carry the operating system and internet exposure the export gives, and --severity-field picks which export column sets each finding's severity
  • Option to have tenable io sync_findings read DISA STIG audit results from the Tenable.io compliance export and send them to the matching control objectives, off by default (tenableIoStigComplianceExport)
  • Vulnerabilities synced by scanner integrations now carry a finding type so non-CVE findings can be told apart from CVEs
  • New wizHeldFindingsQuietSeconds setting (default 30 seconds) sets how long wiz vulnerabilities waits, once no new asset appears, for a concurrent wiz inventory to write the assets of findings it holds

Changed

  • Tanium sync_cis_benchmarks now writes checklists only to assets in the security plan given by --regscale_id or in one of that plan's components; endpoints whose assets are all in other plans or their components are skipped instead of being written to a copy elsewhere, and the run reports how many were skipped
  • A removed init.yaml key the CLI no longer reads (the gRPC ingest keys, cisaKev, jiraDefaultIssueType, jiraDefaultPriority) now logs a warning on every run that names what to use instead, rather than being ignored silently, unless it still holds the value an older CLI wrote into every init.yaml; regscale config validate reports these keys as removed_key
  • Trust Center requests are now refused if they would leave the configured Trust Center host, API errors no longer echo the request's query string, and each request writes one debug log line without headers, bodies or query strings
  • Memory use when importing a large SARIF file for compliance mapping is roughly halved, because the file is now read in a stream instead of being loaded whole
  • Breaking change for custom scanner integrations: the ScannerIntegration flag that turns on closing outdated findings is now named close_outdated_findings_enabled, and the old close_outdated_findings name is no longer accepted as a flag because it is the inherited close method. Passing it to set_keys or to sync_findings/sync_assets on a ScannerIntegration subclass now raises TypeError. Two cases fail silently instead, so audit custom code for them: a class-level or instance-level close_outdated_findings = True assignment overwrites the inherited close method with a bool, and reading close_outdated_findings as a flag (if x.close_outdated_findings:) now returns the method, which is always truthy. Use close_outdated_findings_enabled for both.
  • A refused token now names its likely cause: an expired token says when it expired, and an unexpired one lists the RegScale 6.36 events that revoke tokens (password change or reset, administrator sign-out or deactivation, local sign-in turned off) with the fix for a service account or a person. A refused XCCDF benchmark import now logs the server's status, error code and message, such as the 409 that asks you to delete a benchmark with assessment results before replacing it
  • Nessus import reads large scan files faster and no longer loads the CPE dictionary when a scan names no CPE
  • Scanner imports make fewer requests to RegScale at start-up and spend less time per finding
  • The Nessus import is faster on folders with many files because it frees memory after a set amount of input instead of after every file
  • Trivy and Grype findings no longer start with the CVE in their title, and carry more of the scan data where the scan provides them: references, layers, CWE ids, EPSS and known-exploited flags, fix advice and image details
  • regscale jobs names each run's log file with its start time first, in UTC (20261108T070012Z-<job>-<run id>.log), so a plain ls of job_logs/ lists runs in the order they started. Logs from earlier runs keep their old names
  • Databricks sync_findings --mop-up now closes CIS checks and CVE findings separately, each under its own source (Databricks CIS Benchmark and Databricks CVE), so an export of one type no longer closes the other type's records, sync_assets --mop-up is refused unless the export holds both types, because the two types share hosts, and a plan that already holds issues under the earlier source Databricks gets a second set of issues on its first run, so close or delete the old ones first
  • Synqly connector prompts hide any value that the Synqly schema marks as a password, such as a client secret
  • Aqua imports use less memory on large reports
  • Each Xray vulnerability is now stored under a key that includes its CVE. The first Xray import after the upgrade creates new vulnerabilities with a new first seen date and closes the old ones, which keep their history.
  • On defender export_resources, -n now means --query_name only, so a script that used -n to skip the upload must use --no_upload instead
  • When a scanner sync stops because its vendor source fails partway through, the findings read before the failure are now sent without closing any findings, and a later error while sending them is logged at ERROR so the vendor failure still sets the exit code
  • If all rules of a control are in the conformance pack, aws sync_config_compliance --conformance_pack_name now closes its fixed findings and issues

Fixed

  • The interactive terminal GUI no longer fills the console with random characters when a file upload fails and the server returns a non-text error response
  • jcamValueMapping accepts @drop for a value, declining it: the value writes nothing and leaves the destination as it is, instead of being written literally as the text @drop; a value mapped to @unmapped still works and is noted once per run in favour of @drop
  • jcam import_ssp no longer stops for every plan when JCAM returns an undated SORN notice beside dated ones; undated notices go in the last slots, and an unexpected error in one plan's privacy data now fails only that plan
  • jcam import_ssp now sets each child plan's parent from JCAM's FISMA rollup, including a first parent and a change of parent, instead of every change being refused, and leaves the parent alone when JCAM's parent system has no RegScale plan instead of writing 0
  • jcam import_ssp no longer creates a plan a second time, without its compliance setting, in the front-matter step after RegScale refused to create it; the system is skipped and the refusal stays recorded
  • regscale jcam import_parameters counts a failed read of a plan's controls, or of a control's parameters, as a failure in the run summary, instead of treating the plan as having no controls or the control as having no parameters
  • jcam import_assessments and import_ssp count each assessment or system role RegScale's batch create did not return as a failure, instead of counting a refused batch or item as written
  • tenable nessus import_nessus now sets each asset's type from the host instead of creating every asset as "Other": an asset-type host tag (any asset type value, case and spacing ignored) wins, otherwise the standard system-type property is mapped (router, switch, firewall, load-balancer, hypervisor, mobile, general-purpose, and now embedded and printer as Appliance). The asset category follows the type (Software for application, database, container and cloud types, Hardware otherwise), assets created for findings use the same type, and a host with no type information no longer resets the type already stored for it when a newer scan is imported (its category is still sent as Hardware) (requires RegScale platform 6.35.1.0 or newer)
  • jcam import_poam matches an existing POA&M only among the importing plan's issues, instead of taking any issue in the tenant with the same otherIdentifier and moving it onto the plan
  • Tanium Cloud CIS benchmark checklists now show the rule title and the last scan date instead of only the rule ID
  • ServiceNow incident sync now copies attachments between incidents and RegScale issues instead of always reporting zero
  • AWS asset console links now open the AWS GovCloud console for resources in GovCloud regions
  • AWS Lambda layer version assets are now named by their layer and version instead of the bare version number
  • Jira evidence sync lists the Jira Tasks named in a repeated-failure log line in sorted order, so the line reads the same on every run
  • axonius_v2 commands give up on an Axonius request that times out after one attempt instead of four, and exit 6 with one timestamped log line naming the endpoint, the limit and axoniusTimeout, instead of a traceback or an unlogged error; the connect wait is capped at 30 seconds, as for other vendors, and axoniusTimeout is the read limit
  • jcam import_ssp skips JCAM's placeholder information type 'Other: Other' unless the tenant has a type of that name, instead of counting a plan that lists only the placeholder as failed and offering it in missing_info_types.json
  • The final step of a long import, where the RegScale server finishes the sync run, now logs when it starts, every 30 seconds while it runs and when it is done, instead of going silent for minutes
  • Interactive imports that send many batches, such as a Nessus folder import, no longer fill the terminal with finished sync progress bars that pushed the log output off screen part-way through the run
  • Trivy and Grype findings are dated from their scan file instead of the import time, and with import_scans first seen and last seen no longer move the wrong way when a newer or older scan is imported
  • --s3-bucket and --s3-prefix now read the bucket for Trivy and Grype imports, using the AWS profile or keys given on the command line
  • Compliance integrations, including AWS Security Hub compliance and Wiz compliance_report, count a compliance item that fails to process as a partial failure, so the run exits 3 under REGSCALE_STRICT_EXIT_CODES=1 instead of 0
  • Commands run without a terminal (regscale jobs run-due, cron, a redirected shell) now log a progress line at INFO at most once a minute during long loops, and a closing line with the count and elapsed time, where they used to stay silent until the loop ended
  • regscale jcam import_controls creates a plan's new control implementations 100 per request instead of one request each, and sends a control the report repeats once
  • Databricks sync_findings and sync_assets with --regscale-ssp-id use less memory on a large export, because they no longer keep one object for every distinct host of the whole file while they stream it
  • Tenable.io sync_compliance_controls no longer stops on a compliance check with a null see_also, description or other descriptive field, and skips a check it cannot read instead of ending the run
  • FedRAMP Appendix A import no longer stops when a control in the document has no implementation status selected
  • The FedRAMP POA&M Excel import now accepts workbooks built from template v3.x. It no longer requires the removed Resources Required, Planned Milestones, Milestone Changes and Operational Requirement columns, and it reads Remediation Plan in place of Overall Remediation Plan. It skips a POA&M sheet that has no POAM ID header, with a warning, instead of stopping. Typing a quoted answer such as 'skip' at the mapping prompt no longer loops.
  • Trivy, Grype, SARIF, OpenText, Qualys Total Cloud and Tenable SC file imports no longer fail with "cannot pickle '_queue.SimpleQueue' object" when the staged findings file is larger than 500 MB
  • Compliance syncs now fill in the control and security plan on each control assessment, and the control and assessment on issues whose finding maps to a single control. A new issue gets its assessment on the first sync and its control on the next
  • Critical RegScale issues synced to Salesforce now create cases with High priority instead of Low
  • JCAM milestone import reports an unrecognized or undated completed status instead of guessing, and export sends a milestone's own status to JCAM when it is one JCAM uses instead of always Completed or Not Started
  • JCAM milestone import no longer gives a milestone JCAM sent no due date the import time; RegScale sets the date from the POA&M's severity
  • Reminder emails no longer fail for users who have no upcoming or overdue items, and the run now reports when nobody has any
  • Qualys scanner no longer fails on a finding with no detection data and now gives such findings a unique id
  • regscale tenable io sync_compliance_controls now creates a control implementation for each matched control that passed or failed a check and that the security plan does not have, gives an enhancement control (for example AC-2(1)) the status Tenable.io reported, and exits 1 when it matches controls but RegScale saves none of them
  • The stored size of an uploaded file, such as a Jira attachment or evidence file, is now its byte count
  • AWS sync_assets on plans with assets in more than one region now marks a deleted resource as not in the latest scan and keeps edits made to it during the run, and the AWS compliance commands keep an Inactive status a user set and give S3 buckets the Cloud Storage type without replacing their description
  • On platforms before 6.36.0.0, regscale aws sync_config_compliance now closes the issues that CLI 6.x stored under the bare AWS account ID when it retires those records, instead of leaving them open, with issueCreation: Consolidated (the default)
  • regscale aws sync_config_compliance --conformance_pack_name no longer closes the findings and issues of controls that pass, so the findings and issues of failing rules outside the conformance pack stay open
  • Scan History Date Scanned now shows the scan date instead of the import time for integrations that import vulnerabilities in batches
  • Asset syncs no longer log a duplicate asset that RegScale saved from a later row of the same batch as "not saved", and the "Successfully submitted" and "asset(s) synced" counts now show distinct assets
  • Synqly connector syncs create their Synqly account and integration with the installed Synqly SDK
  • Amazon Inspector syncs through Synqly build an AWS access key credential and ask for the access key ID and the secret access key, instead of stopping with an attribute error
  • Synqly ticketing syncs create tickets and upload attachments with the installed Synqly SDK
  • The error that follows a rejected Synqly integration setup no longer shows the credential values that Synqly echoes back
  • Synqly vulnerability syncs such as Amazon Inspector, Wiz and Tenable Security Center no longer stop with a filter validation error before they fetch findings
  • The file comparison command now names the supported file types, CSV, XLSX or XLS, when it refuses a pair of files
  • If another import saves the same plan, an asset import no longer adds a second copy of an asset. It reports the asset as not saved, with the reason
  • The FedRAMP OSCAL SSP import now ends with one clear error line and exit code 6, or 4 when the host refuses access, when the host of the profile that the SSP names cannot be reached or answers with an HTTP error, instead of a traceback
  • The FedRAMP OSCAL SSP import now reads a profile that its host sends compressed, such as the FedRAMP profiles published on GitHub, instead of failing with a decoding error
  • The FedRAMP OSCAL SSP import now ends with one clear error line and exit code 1, instead of a traceback, when the profile that the SSP names is not a readable JSON document, and it no longer leaves a copy of the profile in the system temporary directory
  • The eMASS POA&M workbook import now stores each POA&M with an Issue status the tenant offers (Ongoing as Open, Completed as Closed with its completion date, Risk Accepted as Exception/Waiver or Open, and Open or Closed as Ongoing or Completed on a tenant that uses the eMASS statuses) instead of losing rows.
  • Tenable.io commands now honor the sslVerify setting, and Tenable.io and Tenable.sc commands now verify against the CA bundle set in customCaCert
  • Tenable.io asset sync stores an asset's MAC addresses in the same order on every run
  • tenable io info shows the Tenable.io user, account email and scanners instead of always reporting a connection error, and exits with a non-zero code when it cannot get them
  • regscale tenable sc list_tags now lists the Tenable.sc tags. It asked pyTenable for a tags endpoint that does not exist, logged an error and exited 0. It now reads GET /rest/tag, prints "No tags found in Tenable.sc." when there are none, and exits 4 when Tenable.sc refuses the keys, 6 when it does not answer or reports an error, and 1 when the reply is not a list of tags.
  • Prisma CSV import removes repeated rows before importing, so the logged counts are of distinct findings
  • Wiz add_report_evidence now fails with an error naming the refused upload, instead of logging that the file was uploaded
  • Jira commands that use basic authentication now honor the sslVerify setting, and a blank sslVerify key verifies the certificate
  • OpenSCAP imports no longer leave two open issues for each failed rule, and the one issue that remains is linked to the vulnerability but not to the affected controls
  • axonius_v2 commands print one error line that names the Axonius host and the failure, and exit with code 1, instead of a traceback when Axonius cannot be reached, drops the connection or answers with an error or an unreadable reply; a timeout still exits with code 6
  • axonius_v2 now prints one error line that names the input, and writes nothing, when a saved query config file is missing or invalid, a saved query is not in Axonius, a regscaleSspId is not an integer, or --offset or --limit is negative
  • Axonius asset sync no longer imports nothing when one device has no preferred hostname or name, or one flat-file row has no CDM Unique ID; such a device is synced under another identifier, and a device with no identifier at all is skipped with a warning that names it
  • Fixed axonius sync_assets --regscale_ssp_id to stop with an error when the security plan does not exist instead of writing assets under it
  • Axonius asset sync now stops with an error and a non-zero exit code when RegScale refuses an asset batch, for example when the token has expired, instead of reporting completion
  • Axonius pull_data_axonius gives up when Axonius sends nothing for the axoniusTimeout setting (exit 6) instead of a fixed 3600 seconds, and reports a refused connection, an unknown host or a reply that is not a JSON object as one error line naming the host (exit 1) instead of a traceback
  • Axonius pull_data downloads only the S3 object that the key names and fails when no object, or several objects, match the key
  • axonius combine_data_parts merges a part that holds one JSON object as one record instead of writing its field names, and stops with a usage error for a part that is not a list or object
  • axonius file commands and combine_data_parts report a missing, unreadable, non-UTF-8, empty, non-JSON or non-list input file, and a tenant without the "FISMA Id" security plan field, as one Error: line with exit code 1 instead of a traceback; the sync commands now check the input file before they query RegScale
  • Report a missing security plan in the Axonius V2 sync commands as one error line instead of a traceback
  • prisma sync_sbom (and panw sync_sbom) now fails when the SBOM download fails or the archive cannot be read, with exit 4 when Prisma Cloud refuses the credentials and exit 6 otherwise, instead of reporting success and exiting 0
  • The Prisma Cloud and Palo Alto sync_sbom commands now accept --no-bulk-download to sync the SBOM of one host or image
  • Prisma authenticate (and panw authenticate) now stops with exit status 4 when the login reply holds no usable token and with exit status 6 when the reply is not JSON, caches nothing, and ignores a tokenless session that an older release cached
  • Prisma Cloud syncs now wait for each Retry-After value, at most 30 seconds, log every retry wait and fail at once when the console asks for a longer wait
  • Burp Suite imports no longer stop when an issue has no background or a request and response pair has no request
  • Importing a Burp Suite export now maps a vulnerability to each host that reports an issue type, where only the first host got it; the issue names every such host, and the stored vulnerability key does not change
  • The Burp import now names a file that is empty, cut short or not a Burp export, and exits with an error when no file in the folder could be imported
  • Burp import no longer prints a time zone warning for an export time that ends in a zone abbreviation such as EDT
  • The Acunetix and Prisma CSV imports, and the other flat-file imports, now print one line for each file they cannot read and exit 1 when no file in the folder could be imported
  • Importing a JFrog Xray report now stores one vulnerability for each CVE that a finding lists, where it stored only the first
  • Fix the JFrog Xray import storing a finding as Low when its CVE has no CVSS v3 score: it now uses the severity Xray gives the finding, or the band of the CVSS v2 score
  • Fix the JFrog Xray import skipping a finding that has no CVE identifier. The import now stores such a finding under its Xray issue id. A warning now names each finding that the import skips
  • Importing a JFrog Xray or AWS Inspector JSON report with no findings now completes and says the report holds no findings instead of asking for a column mapping or failing, and an Xray report page that lists none of its rows is skipped so it closes no findings
  • JFrog Xray import of a report that holds one finding as a single object instead of a list
  • The Xray, Snyk and other JSON imports now name a .json file that is empty, cut short, not valid JSON, not readable or not a list of records, skip it, and import the other reports in the folder. A JSON report that starts with a UTF-8 byte-order mark is now read
  • Prisma CSV import keeps findings identified by a Prisma Cloud or GitHub advisory ID instead of skipping them
  • Prisma CSV import stores a finding with no CVSS score, or a score of 0.0, at the severity in its Severity column instead of Low
  • A mapping file that is empty, not UTF-8 text, not valid JSON or not a JSON object now stops show_mapping with a message that names the file, instead of a traceback or a printed value
  • Scanner file imports and other commands that check their security plan or component now stop with an authentication error when RegScale refuses the account (exit status 4, or 78 when the token cannot be renewed), instead of reporting an invalid plan and exiting 0
  • Defender commands now report that Azure refused the credentials, with Azure's error description, instead of a missing access_token key and a traceback; the exit code stays 1
  • Defender export_resources now reports the HTTP status and the request that failed when Azure refuses or fails the saved query request, instead of a traceback
  • defender export_resources -n "<name>" now exports the named saved query and no longer prints five duplicate-option warnings
  • Microsoft Defender alerts are no longer skipped when an activity time cannot be reformatted, and a time with an offset now shows its UTC date
  • Entra evidence collection logs one error line for each evidence type that Microsoft Graph refuses, in place of a traceback and a bare exit code
  • regscale emass_api authenticate, push_poam and sync_system now stop after one wait when eMASS does not answer, print the line eMASS at <host> did not answer in time (<n> seconds). and exit 6. push_poam no longer sends a new POA&M when its search for an existing POA&M gets no answer, a bidirectional sync_poams prints how many POA&Ms it pushed before the timeout, and authenticate no longer exits 0 when its connection test fails
  • regscale emass_api commands no longer wait without a limit when eMASS accepts the connection and does not answer. The CLI now uses the emass.timeout setting (or EMASS_API_TIMEOUT, default 300 seconds) for each eMASS request. After that time the command prints one line, eMASS at <host> did not answer in time (<n> seconds)., and exits 6. When the request was a change, the line adds eMASS may have applied the change., so check eMASS before you run the command again. A command that sends many items, such as sync_controls or upload_artifacts_bulk, stops at the first timeout. The time limit applies to each phase of a request, not to the whole request, so a server that sends a little data at a time can still keep a command waiting. A large artifact upload to a slow eMASS server can now stop at 300 seconds; raise emass.timeout for it. A value of emass.timeout that is not a positive number, such as 0, -5 or abc, uses 300 seconds and logs a warning.
  • regscale emass_api sync_controls no longer exits 0 when no control synced because eMASS could not be reached or refused the request. When no control synced and at least one control failed because of eMASS, the command prints one error line with the eMASS host and exits 4 if eMASS answered 401 Unauthorized or 403 Forbidden for any control, and 6 for an unreachable host, a TLS failure, any other error status or an answer it cannot read. It used to print [OK] Control sync complete with every control counted as an error. When eMASS refuses only the content of controls (400 Bad Request, 409 Conflict, 422 Unprocessable Content), or when some controls synced, the failed controls stay partial failures: the command prints how many controls were not synced, with the synced and error counts, and no longer prints [OK]. The DEBUG log message for a failed control push no longer carries the eMASS response body.
  • regscale emass_api upload_artifacts_bulk and add_milestones_bulk no longer print tracebacks or the eMASS response headers and body when eMASS fails, sync_poams prints one error line that names the eMASS host when its pull fails, and import_xml prints one error line without a traceback for a file that is not XML
  • The regscale emass_api commands now print one error line that names the problem and the eMASS host, for example eMASS at emass.example.mil rejected the API key: 401 Unauthorized., instead of one to three tracebacks. emass_api test_connection no longer ends with an empty Error: line and Aborted!. The line and the warnings never carry the API key, the response headers or the response body. A 401 Unauthorized or 403 Forbidden answer exits 4. An unreachable host, a failed TLS connection, any other HTTP error status or an unreadable answer exits 6. Both used to exit 1. A local or RegScale error still exits 1.
  • The regscale emass_api commands delete_milestones, add_milestones_bulk and upload_artifacts_bulk accept --yes (or -y) to skip the confirmation question, so a scheduler, a pipeline or the Orchestration Hub can run them. Without --yes and without a terminal, each command now prints one Error: line that names --yes and exits 2 before it calls eMASS, instead of a traceback and Aborted!.
  • regscale emass import_slcm now reports a file that is empty or is not an Excel workbook as one error line that names the file, and exits 1, instead of printing a traceback
  • The eMASS import commands and detect_workbook now report a file that they cannot read as one error line that names the file and exit 1, instead of a traceback or exit 0
  • regscale emass import_slcm now exits 1 when the security plan or the catalogue is not found, 4 when RegScale refuses the lookup and 6 when RegScale does not answer it, instead of exiting 0 after logging "Aborting!"
  • If a control has the Implementation Status Manually Inherited, a blank status or an unknown status, the regscale emass import_slcm command no longer marks it Not Applicable. Manually Inherited is now saved as Manually Inherited, as regscale emass import_slcm_workbook does. The Implementation Status is matched without regard to case. A blank or unknown status keeps the control status and its N/A Justification, and logs a warning that names the control. The command applies Not Applicable from the Implementation Status or Compliance Status column only with an N/A Justification. Without one, the control keeps its status and the command logs a warning.
  • The eMASS hardware and software inventory import now reports each row it skips and counts each asset RegScale refuses as an error
  • eMASS POA&M workbook import now stops and names the rows that share a POA&M Item ID, instead of merging them into one issue
  • eMASS control test results import now names the row and value of an unreadable Date Tested when the CCI does not resolve
  • eMASS control test results import no longer sends an unreadable Date Tested to RegScale and names the CCI number in its errors
  • OpenSCAP import_csv now names a CSV file that is cut off, has a header that matches no supported layout, is not UTF-8 or has a field over the CSV size limit, and exits non-zero instead of importing part of the file or nothing
  • OpenSCAP imports now name each missing, empty, truncated or non-XML scan file, and a folder with no scan file, and exit non-zero, while the other files in the folder are still imported
  • OpenSCAP CSV import now reads files that start with a UTF-8 byte-order mark, such as files saved by Excel as CSV UTF-8, and imports the same findings as the file without the mark
  • cci_importer prints one error line that names the file, with no traceback, when the CCI file is empty, truncated, not XML or cannot be read
  • regscale stig cci_assessment and regscale stig_mapper run now stop with an error and a non-zero exit status when the security plan does not exist, in place of reporting that they found nothing to do
  • The jira_dc commands now log one error line that names the problem and the Jira Data Center URL, with exit code 4, 5 or 6, when the first request to Jira Data Center fails, and a server that does not answer fails after jiraDcTimeoutSeconds instead of about six minutes
  • Jira commands jira issues and jira tasks now end with one error line and a non-zero exit code when Jira refuses the credentials, cannot be reached or answers with an error, and they report an unset jiraUrl, instead of printing a traceback
  • Qualys scan import no longer stops when a row's severity is empty, not a number or outside 1 to 5; it logs a warning naming the row and imports that row at low severity
  • The example Qualys URL in the init.yaml template is now a host only, and an existing init.yaml that keeps the old example path no longer doubles it or triggers config validation warnings
  • Qualys diagnostics and CSV validation now exit with a non-zero status when a check fails, and a failed request no longer prints a traceback
  • The Qualys import_policy command now reports a one-line error and exits with status 2 when the file is not a Qualys policy export or an option value is not valid, and no longer creates an Untitled Policy from a JSON object that has no policy ID
  • Qualys import-vm-report now stops with a usage error when both --report-file and --report-title are given, instead of silently importing the file and ignoring the title
  • A mapping file that is not valid JSON, not UTF-8 text or of the wrong shape now stops the import with exit status 2 and one line that names the file, instead of a traceback
  • Import a Qualys VM scan row that has no QG Host ID as an asset keyed on its IP address, and skip a row with neither
  • A Wiz token URL that ends in any mix of slashes and whitespace, such as a space or a tab followed by a slash, is now accepted, as one that ends in a slash or a space already was
  • Wiz commands now name the token endpoint when its answer is not a token, instead of a redacted placeholder
  • regscale wiz attach_sbom now attaches an SBOM whose document has no version, and when RegScale refuses one SBOM it still attaches the others and then exits 1
  • A compliance sync that stops on a usage error, such as a Wiz compliance report that is not UTF-8 text, prints one "Error:" line and no longer also logs the same message with a traceback
  • Wiz inventory, issues and vulnerabilities help now names the --filter_by_override option and shows an example that is valid JSON
  • The deprecated wiz sync_compliance command no longer lists --list-frameworks in its help, and using that option now stops with a usage error (exit code 2) that points to wiz compliance_report instead of exiting 0 after printing a message
  • The tenable was sync_findings command no longer exits 1 before it sends a request to Tenable.io when the security plan already holds assets
  • A Tenable.io --tags value without key:value pairs, an --asset-filter-file that is not a JSON object, and a Tenable.sc sync_jsonl --mapping_file that is not a JSON object now end with a usage error that names the option and the expected format. A file passed as a pipe, such as a shell process substitution, still works
  • Tenable.io CIS checklist syncs and Tenable.io STIG compliance results now send asset tag filters in the form Tenable.io accepts, so a run with --tags no longer fails and syncs nothing
  • Tenable.sc file import with --file-path now stops with an error and a non-zero exit status when no file in the folder can be read as Tenable.sc data
  • Negative --offset and --limit values are now refused with a usage error instead of stopping a sync or processing the wrong records
  • Log redaction no longer hides the start of a connection error after a sign-in URL that ends in /token, so "[Errno 111] Connection refused" stays readable
  • When Microsoft Defender keeps failing or does not answer, the error now names the request URL again, without its query string or any credential
  • A request that keeps failing with an HTTP error such as 500 Internal Server Error or 429 Too Many Requests now logs one error without a Python traceback, and Microsoft Defender reports a sustained failure once
  • A missing required setting, such as wizUrl, now ends a command with one error line and exit code 5 instead of a Python traceback
  • tenable io sync_compliance_controls keeps the control owner of each control implementation it updates
  • Qualys import-vm-report now saves Re-Opened and Fixed detections with a RegScale vulnerability status, and exits non-zero when it saves no asset or vulnerability or a request fails
  • The eMASS asset, POA&M and test result integration classes now read the eMASS responses and sync the systems, POA&Ms and test results they return instead of failing with an error
  • Tanium benchmark and ports syncs stop with an error when RegScale refuses or fails the read of the plan's components, instead of skipping the assets under those components
  • Tanium ports and protocols sync in an Orchestration Hub split job now reads ports for the same endpoints that the job's asset step writes
  • Tanium ports and protocols sync now writes ports only to the endpoint's asset in the selected plan or its components, not to copies of the endpoint in other or deleted plans
  • eMASS and FedRAMP imports that update control implementations no longer reset the inherited record permissions or clear the last assessment, the parent control and the approval state of those control implementations
  • If two rows of an asset sync differ only in letter case, the sync writes the software and SBOM of their asset once
  • The jira_dc sync_tasks command no longer creates two RegScale tasks when the Jira search returns the same ticket on more than one result page
  • FedRAMP OSCAL SSP imports now store the ports and protocols of each component, which RegScale rejected before because their Service and Used By values were empty
  • Compliance syncs on RegScale 6.36+ read each control's newest same-day assessment from POST /api/v3/assessments/byParents, so a control assessed more than once today merges with its latest result instead of its earliest.
  • AWS sync_assets now shows the valid from date, the valid until date and the number of resources in use on each ACM certificate asset
  • Qualys sync keeps a consolidated issue and vulnerability open while any host or web application still reports the vulnerability as open
  • Qualys WAS findings that Qualys reports as fixed now close their issue and vulnerability during sync
  • env_info now prints its full report in the container image, under an unknown user id and in a read-only directory
  • Pipeline report_scan no longer stops before it creates the issues for a failed scan when the SSP-level control update fails, and RegScale no longer refuses that update because the control origination is missing
  • Pipeline setup_component and report_scan now link the repository component to the security plan, so it appears in the plan's Components tab
  • regscale wiz vulnerabilities now names every finding it did not write when it exits 1 because regscale wiz inventory had not written the plan, one line for each finding with its asset
  • aws sync_findings without --generate-evidence or --format now sends the findings it read before a failed Security Hub page, instead of none of them
  • A scanner sync that the server did not start, because another sync of the same source was still running, no longer sends its checklist results or changes the status of a control or control objective
  • Checklist roll-ups send only valid assessed dates, and a Tenable STIG result whose scan date is missing or not a valid date is dated by its last seen date, so one bad date no longer stops every result in the sync from updating control status
  • Tenable STIG imports send a passing check's vulnerability as closed and any other result as open, instead of sending the check result, which the server rejected with a warning for every passing check
  • Wiz vulnerabilities and issues incremental syncs run with --offset or --limit no longer move the saved pull watermark past findings they did not write
  • Stop logging an error and stack trace when the catalog update cannot read the platform feature flags
  • A Tenable.io asset or vulnerability sync now stops before it sends any record from a cached export that is corrupt, and an export that stops partway is no longer cached for the next run
  • A truncated Tenable.io asset or vulnerability export cache is now removed and fetched again on the next run instead of failing every run until it expires
  • Updated the pyasn1 dependency to a release that fixes known security vulnerabilities
  • BigQuery asset sync now reports a rejected query as a single error line instead of a Python traceback
  • Issues from a component compliance report are now saved under the component and no longer rejected as a missing security plan
  • A Qualys CSV import whose column changes type after the first 10,000 rows now stops with a one-line message that names the file instead of a traceback
  • A Qualys file import with a negative --skip_rows value, or one that reads a title row as the column headers, now stops with a message that names --skip_rows instead of a traceback
  • Tenable compliance sync gives a failing control its planned implementation date and steps to implement, and sets a passing control to Fully Implemented, on platforms that take bulk control updates
  • Qualys import_total_cloud no longer logs that it imported findings RegScale refused, and reports how many vulnerabilities RegScale stored
  • Re-importing a Qualys VMDR report into a security plan now updates each host's existing asset instead of creating a second asset for every host
  • The eMASS POA&M workbook import now stores the controls, devices affected, impact description and security plan of each POA&M it creates, and fills the false positive, operational requirement, auto-approved and risk adjustment flags with No when they are empty. A row the batch upload refuses, such as a Risk Accepted row stored as Exception/Waiver, is now written on its own, or reported with the server's reason, instead of being dropped with "see the log".
  • FedRAMP CIS/CRM imports store every control and part inheritance model when the RegScale platform does not accept them all in one bulk request, and fail with an error line when RegScale refuses a model
  • Evidence and scan results now map to every control of a security plan with more than 500 control implementations when RegScale leaves out the next page marker on a full page
  • Evidence and scan results now map to control enhancements on a RegScale platform with the bulk compliance routes, and evidence also maps when the source names a control enhancement in the AC-2(1) form
  • On RegScale 6.36 and later, a compliance sync no longer replaces the result of an assessment from another day
  • Control implementation syncs and imports on RegScale 6.36.0.0 or later now retry or report a control implementation that the bulk update did not write, instead of finishing as if it succeeded
  • Wiz syncs that run longer than the Wiz access token lasts get a new token and keep going, instead of failing with an HTTP 401 and exit code 4

FedRAMP authorization carries ongoing reporting obligations: vulnerability reports, significant change notifications, incident reports, and recurring certification reports, each with its own deadline. The data behind them is sensitive, so providers need a secure system of record for managing these obligations and precise control over who can see what.

RegScale Trust Center 1.0.0 is now available in Beta. For cloud service providers (CSPs) pursuing or maintaining FedRAMP Rev 5 or 20x authorization, Trust Center provides a secure place to manage reporting obligations and share the appropriate view of their authorization data with assessors, customers, and partners.

Availability

Trust Center 1.0.0 is ready for production data, so you can use it with your live FedRAMP information. The Beta designation reflects that many features are in minimum viable product (MVP) form: they support the core workflow from start to finish, with less depth, configuration, and refinement than we plan for general availability. Screens, defaults, and workflows may change before general availability.

Core Capabilities

KSI Management

  • Record the status, narrative, and evidence of each Key Security Indicator (KSI), with completion rolled up by family.
  • Select pre-configured measurement methods from Automation Packs, or write your own. AI-assisted drafting is available when AI is turned on for your organization.
  • Report on completed KSI's in the Trust Portal or Auditor Portal.

FRR Management

  • Under the FedRAMP Consolidated Rules for 2026 (CR26), set each FedRAMP Rule (FRR) that applies to your certification class to Compliant, Partial, Non-Compliant, or Not Applicable, and add a narrative and evidence.
  • Walkthroughs in the application explain each rule family.

VER/VDR Reporting

  • Evaluate vulnerabilities in Core Platform, including each finding's potential agency impact (N1 to N5) rating. Trust Center generates the Vulnerability Evaluation and Reporting (VER) and Vulnerability Detection and Response (VDR) reports and artifacts from that analysis. Track each reporting obligation with its next due date and status.
  • Choose how findings reach Trust Center. Route scanner data through the RegScale Orchestration Hub (ROH) into Core Platform and sync it to Trust Center, or load findings, with their evaluations, directly into Trust Center through its API.
  • Publish the monthly Vulnerability Detail, Accepted Vulnerability Information, and overview reports as a single package, validated against FedRAMP's schemas. Preview each publish before it goes out, and download the published month as a PDF. Historical VER Activity is available as a separate, all-time report.

Rev 5 Artifact Management

  • Import an existing Rev 5 Word SSP, its Appendix A, or an OSCAL JSON SSP. Then maintain control implementations, inventory, POA&Ms, the Customer Responsibility Matrix, and five narrative appendices through guided wizards, with completion tracked in one view.
  • Publish signed OSCAL files and a human-readable SSP from the same source data. A guided workflow supports migrating an existing Rev 5 authorization into Trust Center.

Trust Portal

  • Registration is self-service or by invitation, with an optional NDA. Visitors request document access with a justification. Your team approves with view-only or view-and-download access, or domain rules decide by email domain.
  • Share a system that is not yet visible to everyone with one named viewer, such as a pilot customer.
  • Choose which of the portal's 13 tabs appear, and in what order.

Auditor Portal

  • Invite an assessor firm to audit a system with a defined scope and access window. Assessors see only the systems you enable for them.
  • Assessors record Met or Not met verdicts, raise findings, and request and review evidence.

Customizable UI

  • Use the CR26 Guided Tours to educate your users on how requirements align with capabilities in the application.
  • Use the Hide Features settings to turn off any sidebar or profile-menu items for users in your organization. New tenants start with a core capabilities feature set, with the option to enable additional features at any time.
  • Each FedRAMP system has its own Hide Features setting; new systems start with a focused core set.
  • Apply your branding to the portal for a cohesive user experience.

[7.0.1] - 2026-11-04

No changes to CLI commands or behavior from 7.0.0

Risk Management

Enterprise Risk Rollup: The Risk Rollup status board has been redesigned to allow for a single, scoped view of enterprise risk. Set the scope to consistently apply filters across every view in the board, so risk leads can better evaluate where risk is concentrated, and which treatments deliver the most value for the budget. Enhancements include:

  • Scope bar and KPI cards: A new scope bar filters every view, and six new KPI cards summarize scoped risks, risk appetite, KRI health and annual loss expectancy.
  • New risk views added: Three new tabs have been added to the status board:
    • By Lens: A spider chart of scores by impact area
    • Concentration Matrix: In-scope risks presented in a two-dimensional heatmap
    • Business Case & ROI: Risk treatments ranked by return on security investment against your budget
  • Risk inspector: Open any risk row in a slide-over inspector, enabling you to preview a risk's details without leaving the board.
  • More accurate totals: Closed and Cancelled risks are excluded by default, risks recorded on a system now roll up to its organization, and results respect each user's app and record access.

Security Plans

Streamlined Setup, Import and Export: This release streamlines how security plans are created, imported, exported and maintained.

  • Setup Wizard Improvements: The RMF path of the security plan wizard offers an optional Categorization step and now automatically assigns the Wayfinder of the chosen compliance setting.
  • Bulk Profile Import: The Profile Importer now accepts several security profile files at once, so you no longer need to repeat the import for each profile.
  • Navy eMASS Exports: Security plan exports for eMASS now follow eMASS's current designated formats, making it easier to submit your security plan data to eMASS without rework.
    • SCF export: Matches the USN SCF v1.6 template.
    • SLCM export: Includes a Control Set column and completed assignment values.
    • System Details exports: New exports for each supported eMASS template.
    • Pre-export validation: Flags any required information missing from the security plan before you download a System Details export.
    • eMASS formatting: System Details exports use eMASS's own wording and formatting, such as multiple selections separated by semicolons.
  • Asset Import from the Excel Template: The Excel template used by the Import Security Plan wizard has been updated to include a dedicated tab for capturing asset details. On import, each asset entry is created and linked to the security plan automatically, so you no longer have to add them one at a time.
  • Large Record Deletion: Deleting large security plans and assessments is now faster and more reliable, with less impact on other users working in the platform while the deletion completes.

Wayfinder

Wayfinder Usability: The Wayfinder side panel and builder have been enhanced to make them easier to read and organize.

  • Simpler side panel: The Wayfinder side panel on a record now shows one progress summary, one status icon per step and one primary action, so users can see where a record stands and what to do next at a glance.
  • Builder: Activities can now be moved to a different stage from the activity's menu, making it quicker to reorganize a Wayfinder as your process changes.

Questionnaires

Questionnaire Date Entry: Date columns in Table questions now open a calendar picker when selected, giving questionnaire respondents a predictable, consistent way to choose a date.

User Experience

Navigation and Usability Improvements: A variety of changes have been made to the user interface to improve the user experience and usability. These include but are not limited to the following:

  • Searchable pickers: Dropdown and selection fields with 10 or more options now include a search box, so users can find the value they need by typing instead of scrolling through a long list.
  • Back button: The Back button now takes users to the page they were just on, or up to the parent page if there isn't one, so moving through the platform is more predictable and users don't lose their place.
  • Breadcrumbs: Clicking a record's name in the breadcrumb trail now takes you to that record's main page, so you can return to it from any sub-page in one click.

Bug Fixes and General Maintenance

This release includes security hardening, patching, and fixes that improve reliability, performance, and overall platform stability.