· intuitem · News  · 7 min read

What's New in CISO Assistant — Week 49, 2026 (v4.3.1 – v4.3.4)

v4.3.1 to v4.3.4 bring an experimental auditor pre-read for ISO 27001 certification audits, a GLPI integration, automatic reference IDs, workflow-driven field defaults, implementation, documentation and maturity scores in audit reports, MCP document tools and ChatGPT sign-in, plus ISO 9001, SAMA IT Governance, NDMO, the EU cloud sovereignty frameworks and new CMMC, NIST 800-171 and Essential Eight mappings. Auto-lock of overdue audits is now off by default.

v4.3.1 to v4.3.4 bring an experimental auditor pre-read for ISO 27001 certification audits, a GLPI integration, automatic reference IDs, workflow-driven field defaults, implementation, documentation and maturity scores in audit reports, MCP document tools and ChatGPT sign-in, plus ISO 9001, SAMA IT Governance, NDMO, the EU cloud sovereignty frameworks and new CMMC, NIST 800-171 and Essential Eight mappings. Auto-lock of overdue audits is now off by default.

This post covers v4.3.1 (November 14), v4.3.2 (November 15), v4.3.3 (November 17) and v4.3.4 (November 21). The first three are small patches on top of v4.3.0; v4.3.4 is a large one, with an experimental module to prepare certification audits, a GLPI integration and a batch of new frameworks and mappings. One change in v4.3.4 alters a default, so read the next section before upgrading.

Worth Knowing Before Upgrading

  • Overdue audits are no longer locked automatically — audits from a campaign or a third-party assessment used to be locked and moved to “In review” the night after their due date. This now only happens if an administrator turns on Auto-lock overdue assessments under General settings > Workflows. If you rely on the old behaviour, enable it after upgrading (PR #5314). Thanks to @Mohamed-Hacene.
  • curl leaves the images in the next release — as announced with v4.3.0, update your Docker Compose files to the new health checks (see the example in the repository) if you haven’t already.
  • Keyword search in the AI chat needs the vector collections to be recreated and the setup and indexing commands rerun; until then, library search stays vector-only (PR #5318).

Auditor Pre-Read (Experimental)

A new module helps you get ahead of the certification auditor on ISO/IEC 27001:2022 audits. For each requirement, it shows what the auditor will ask (questions, requested records, sampling, interviewees and common pitfalls, taken from an audit guide), what you can already show, and what to do next. A readiness verdict — gap, weak, to confirm, ready or excluded — is computed from the evidence found through the requirement, its applied controls and its recurring tasks: whether a file is present, its status, whether it expires before the audit, and open findings from earlier audits in the same domain. Evidence can be paired with the auditor’s requests, and each next step is one click: upload, pick existing evidence, or set a request aside with a reason.

The module also holds the audit preparation itself (stage, dates, certification body, auditors, interview roles mapped to actors, session schedule) and a “Prepare this session” mode that walks through requirements one at a time, worst first. Views are available by requirement, by audit session, as a copyable evidence request list and as an interview plan. Audit guides are a new library object, so more standards can follow (PR #5308).

Integrations

GLPI joins Jira and ServiceNow as an ITSM provider. Applied controls sync with GLPI tickets, changes or problems in both directions (name, description, status, priority and ETA), and assets sync with GLPI computers, software, appliances and other standard types. It uses the GLPI 11 high-level API with OAuth2; incoming changes arrive through GLPI’s native webhooks, checked by their HMAC signature, and the item is always read back from GLPI so late or replayed deliveries can’t restore old values (PR #5335). Thanks to @nas-tabchiche, who also fixes a 403 when creating integration configurations and errors in the remote-object pickers (PR #5332).

Workflows and Automation

  • Field defaults — an “On event” trigger can now run immediately, inside the save, so its result is visible on reload. This is how you set your own defaults on new objects: a new template moves evidence created as Draft to In review. Only conditions, variables, logs, date offsets and read/create/update steps are allowed in such a workflow, and a failing run never breaks the user’s save. Status fields on evidence, security exceptions and findings can now be updated by workflows (PR #5312).
  • Requester on validation flows — a workflow can name who requests a validation flow, so that someone can answer when the approver asks for changes. Before, the request was held by the run’s identity and stayed stuck (PR #5309). Thanks to @Claquetteuuuh.
  • Findings on requirements — workflow actions can create findings linked to a requirement assessment, creating or reusing the audit’s findings binder (PR #5321). Thanks to @monsieurswag.
  • Automatic reference IDs — supported objects can get sequential reference IDs with configurable prefixes, numbered within their parent. Turn it on in General settings and backfill missing IDs; existing ones are never changed. Reference IDs show up in more tables and exports, and an exact match comes first in global search (PR #5313).

MCP and AI

  • ChatGPT can sign in to the MCP server: an app asking for scopes it may not have now gets the ones it may have, instead of being refused (PR #5295). The consent page describes local destinations more clearly, and administrators can allow or block AI app connections for several users at once (PR #5298).
  • Document tools — get_documents, create_document and update_document let agents read and write managed documents such as policies and procedures (PR #5248). Thanks to @jledoze.
  • Chat search now covers evidence attachments and published documents, respects folder permissions and stays in sync when they change (PR #5293). Library search combines semantic and keyword matching (PR #5318), and PDF text extraction moves from PyMuPDF to pypdfium2, a permissively licensed library that is about twice as fast (PR #5319).

Audits, Reports and Risk

  • Scores in audit reports — the PDF report, the attestation and the Word report show implementation, documentation and maturity scores per category and per requirement, exactly as the audit computes them, with “not applicable” left out and an Overall row. CSV and Excel exports add the documentation and maturity scores (PR #5322). Thanks to @Mohamed-Hacene.
  • Briefing on the home page — draft and deprecated items no longer count, and the summary gains overdue NIS2 incidents, vulnerabilities and audits, expiring acceptances and exceptions, and evidence and contracts nearing their deadline, grouped by urgency (PR #5316).
  • EBIOS RM imports (Pro) accept the ARM zip of per-table exports as well as full-study Excel files, and pick up control statuses, feared-event links and several attack paths per strategic scenario (PR #5327).
  • Task history shows one row per year and one cell per month, with the whole planning horizon ahead (PR #5294). Thanks to @eric-intuitem.
  • Tasks on third-party audits can be assigned to the vendor’s own representatives, who are prefilled in the form (PR #5354). Thanks to @Claquetteuuuh, who also makes the responsibility matrix actor picker searchable (PR #5323).
  • EU and EEA can be chosen as data locations, with GDPR analytics and DORA exports handling these regions (PR #5326).

Bug Fixes

  • Suggested applied controls now get the translated description, not only the translated name (PR #5338). Welcome to @finsecguy, a first-time contributor.
  • The data wizard reads an implementation or documentation score filled on its own instead of silently skipping the row (PR #5336), and matches imported assets on reference ID and name together (PR #5342).
  • The mapping projection bar counts coverage correctly (PR #5353), the metric quick input honours the date (PR #5299), the Word export includes requirement observations again (PR #5300), and multi-page forms submit on their last page (PR #5305).

Framework & Library Updates

  • ISO 9001:2015 outline — quality management systems, clauses 4 to 10, in English and French (PR #5283). Thanks to @Qnadia.
  • SAMA IT Governance Framework — 240 control requirements across 4 domains, complementing the SAMA CSF and BCM libraries (PR #4387), and the NDMO Data Management and Personal Data Protection Standards from SDAIA — 15 domains, 77 controls and 190 specifications, with priorities as implementation groups (PR #4388). Thanks to @smakarim.
  • EU cloud sovereignty — the European Commission’s 2026 Cloud Sovereignty Framework (48 scored questions, SEAL levels) and the proposed Cloud and AI Development Act (CADA), whose requirements may still change before adoption (PR #5351).
  • New mappings — CMMC 2.0 ↔ NIST SP 800-171 rev2 (PR #5343), NIST SP 800-171 rev2 ↔ rev3 built from NIST’s change analysis (PR #5347), CMMC 2.0 ↔ NIST CSF 2.0 (PR #5362), and ASD Essentials ↔ CSF 2.0 and Essential Eight ↔ SCF (PR #5350).
  • Essential Eight requirements now link to reference controls and threats, with five new threats in the intuitem common catalog (PR #5357).
  • German text for the CRA annexes (official EUR-Lex wording), metric, workflow and threat libraries (PR #5245). Thanks to @hlederhaas.

Localization

  • Hungarian — missing interface strings, email templates, the eleven policy templates, SoA export labels and metric units (PRs #5328, #5352, #5360, #5356, #5341). Thanks to @aafeher.
  • Italian — about 400 new strings and a review of the privacy terms (PRs #5301, #5349). Thanks to @LoZio.

For full details, see the v4.3.1, v4.3.2, v4.3.3 and v4.3.4 release notes on GitHub.

Back to Blog

Related Posts

View All Posts »
What's New in CISO Assistant — Week 48, 2026 (v4.2.3 – v4.3.0)

What's New in CISO Assistant — Week 48, 2026 (v4.2.3 – v4.3.0)

v4.3.0 lets AI apps connect to the MCP server with an OAuth sign-in, adds an evidence review queue with a four-eyes option, a briefing-style home page, a richer framework report with evidence zips, scenario selection across risk assessments and their exports, a denser control plan, federated service accounts, a dedicated space for third parties and optional memcached, plus ITSP.10.033, ISO 14001:2015, LYSS and a Hungarian NIS2 preset.

What's New in CISO Assistant — Week 47, 2026 (v4.2.2)

What's New in CISO Assistant — Week 47, 2026 (v4.2.2)

v4.2.2 adds NIS2 and GDPR incident follow-up with deadline reminders and CCB/APD notification worksheets, cumulative maturity levels and weakest-link scoring for Essential Eight and ASD Essentials, asset lifecycle stages, simpler metrics, per-role field access in audits, EBIOS RM study duplication, a report time zone, and new libraries: ASD ISM September 2026, Swift CSCF 2026 and 2027, SAMA BCM and the German NIS-2 Implementation Act.

What's New in CISO Assistant — Week 46, 2026 (v4.2.0 – v4.2.1)

What's New in CISO Assistant — Week 46, 2026 (v4.2.0 – v4.2.1)

v4.2.0 moves the frontend to SvelteKit 3, adds notifications for assignments, approvals and reviews, Typst templates for PDF audit reports, vendor posture views by tier, required evidence in questionnaires, a portals editor that links documents, EBIOS RM workshop 5 in Excel export, order-independent mapping inference, Norwegian Bokmål and Turkish, and new libraries: SOC-CMM, PIPEDA, Czech Decree 409/2025, NCA DCC and CSCC, OWASP Top 10 2025 and MonServiceSécurisé.

What's New in CISO Assistant — Week 45, 2026 (v4.1.3 – v4.1.5)

What's New in CISO Assistant — Week 45, 2026 (v4.1.3 – v4.1.5)

v4.1.3 to v4.1.5 bring tiering to solutions with propagation to their vendors, an experimental spreadsheet import wizard with column mapping, labels on assessments and studies with a label management view, DPIA import from CNIL PIA exports, a Microsoft Graph mail transport, questionnaire compliance on entity assessments, a rebuilt n8n node, SAML account-linking fixes and the Brotli fix.