· intuitem · News · 4 min read
What's New in CISO Assistant — Week 23, 2026 (v3.17.1 – v3.17.2)
Two releases close out the week: v3.17.1 brings a Prometheus metrics endpoint, user-configurable date formats, expanded comments and audit aggregation; v3.17.2 piles on an expanded AI/MCP server, the ABRO framework, a tables column selector, action-plan cost breakdowns, SSO redirect handling, and a big batch of data-wizard, framework-builder, and ordering fixes.
Two releases this week: v3.17.1 on June 3 and a feature-heavy v3.17.2 on June 5. The theme is operability and depth — observability hooks, a meatier AI/MCP server, a new framework, more flexible tables and assets, smarter SSO, and a generous round of fixes.
Headline Features
- Prometheus metrics endpoint — A new endpoint exposes metrics for Prometheus scraping, making it far easier to monitor a CISO Assistant deployment with standard observability tooling (PR #4061), with a follow-up to expose metrics for enterprise as well (PR #4258). Thanks to @melinoix.
- Expanded AI / MCP server — The MCP server gains vulnerability CRUD, asset classes, and richer risk scenario reads — broadening what AI assistants can do against your GRC data (PR #4256). Thanks to @jledoze.
- Cost breakdown on action plans — Action plans now surface a cost breakdown, bringing budget visibility directly into remediation planning (PR #4272). Thanks to @ab-smith.
- Audit aggregation through ancestors — Audit results can now aggregate up through their ancestor hierarchy, giving cleaner consolidated views across nested scopes (PR #4245). Thanks to @ab-smith.
- Expanded comments — The comments feature has been broadened for richer collaboration across objects (PR #4253). Thanks to @ab-smith.
Framework & Library Updates
- ABRO Framework — A new ABRO framework library is now available (PR #4252). Thanks to @tarkadia.
- Journey preset filter in library list — The library list gains a preset filter for journeys (PR #4279). Thanks to @tarkadia.
- Library tooling —
convert_library_v2.pynow supports Threats annotations (PR #4274), plus an updatedexample_framework.xlsxand restored French metadata in the Vendor Due Diligence framework (PRs #4275, #4276). Thanks to @tarkadia.
Data Wizard
- Header normalization — The data wizard now normalizes headers on import, smoothing out messy source files (PR #4160). Thanks to @martinzerty.
ref_idsupport across the wizard — All data-wizard flows now acceptref_idin addition to name, plus a new test harness to keep the wizard honest (PRs #4087, #4151). Thanks to @tchoumi313.
UX
- Tables column selector — Tables now offer a column selector, letting users tailor which columns are shown (PR #4263). Thanks to @ab-smith.
- User-configurable date format — Users can now choose their preferred date format (PR #4236). Thanks to @ab-smith.
- 1–3 properties scale on assets — Assets can use a 1–3 properties scale (PR #4281). Thanks to @ab-smith.
- Task occurrence description — A task’s occurrence now shows its description (PR #4267). Thanks to @Axxiar.
- Higher autocomplete cap — The autocomplete lazy-mode cap was raised to 20 for more useful suggestions (PR #4268). Thanks to @ab-smith.
Security & Infrastructure
- SSO redirect handling — After SSO login, users are now returned to the URL they originally requested instead of a generic landing page (PR #4255). Thanks to @tchoumi313.
- Allowed IPs list — Infrastructure configuration gains an allowed-IPs list, letting operators restrict access at the application layer (PR #4250). Thanks to @tchoumi313.
- Hide “About” menu for third parties — The About menu is now hidden from third-party users (PR #4240). Thanks to @melinoix.
Performance
- Faster framework library updates — Fixed an O(N²) dedup in the framework library update path (PR #4264). Thanks to @Mohamed-Hacene.
Bug Fixes
- Framework builder nodes at deep nesting — Builder nodes are no longer unusably narrow at deep nesting levels (PR #4271). Thanks to @nas-tabchiche.
- Ordering with negative number values — Fixed wrong ordering for selects with negative number values (PR #3618). Thanks to @monsieurswag.
- Name column ordering — Sorting by the name column now behaves correctly (PR #4242). Thanks to @Mohamed-Hacene.
- Risk-assessment scenario filters — Scenarios can now be filtered by current level and residual level (PR #4257). Thanks to @tchoumi313.
- Incident uniqueness and freshness — Incident uniqueness is now limited to
ref_id(PR #4266), and an incident’s last-update timestamp refreshes when timeline entries change (PR #4280). Thanks to @ab-smith. - Vulnerabilities link in findings — Fixed the link to vulnerabilities from findings (PR #4241). Thanks to @Mohamed-Hacene.
- Markdown rendering and translations — Restored a missing
MarkdownRendereracross several pages (PR #3983) and fixed untranslated framework name and description (PR #4278). Thanks to @tarkadia.
Internationalization
- 🇨🇿 Czech localization — A comprehensive update to the Czech translation (PR #4261). Thanks to @zdenek-pergl.
Maintenance
- Vite 7 upgrade — The frontend moved to Vite 7, with the associated Vitest upgrade (PR #4247). Thanks to @ab-smith.
- Pull request template — A PR template was added to streamline contributions (PR #4243). Thanks to @nas-tabchiche.
- Documentation touch-ups — Updated CyFun export docs and fixed a broken link for ENISA’s risk acceptance (PRs #4238, #4244).
For full details, check out the v3.17.1 and v3.17.2 release notes on GitHub.