Changelog | CyfroSec Documentation

Version 1.4 (August 2026)

Encryption Management: added a new administrative workbench to monitor encryption status and manage security evidence.

Data Encryption: enhanced platform security by encrypting sensitive customer data at rest across billing, identity, and code security modules.

Website cookie consent: the public website now has a cookie consent banner with per-category controls. Analytics load only after you opt in, and you can change your choice anytime from the footer.
Bitbucket Cloud: connect Bitbucket repositories for scanning and AI-assisted fixes, alongside GitHub.
Guided remediation: when an automatic fix can't be generated safely, you now get clear step-by-step remediation guidance instead of a dead end.
Re-run a fix: regenerate a patch attempt while keeping the history of previous drafts.

Announcement scheduling: announcement and maintenance banners now support one-time or recurring windows, with pause, resume, and quick-extend controls, plus a delivery history of every publish.
Website footer: now shows the company registration details and quick links to the Privacy Policy, Terms of Service, and Contact.
Clearer AI assistant status, with visible "Thinking…" and "Working…" indicators while it processes.
Generated fixes are now restricted to approved files only, for safer automated patching.

Announcement banners now reliably disappear once their scheduled maintenance window ends.

Announcements & maintenance pages: admins can publish a site-wide announcement banner or a full maintenance page, with a draft-then-publish workflow.
Business-logic analysis: detects logic-level vulnerabilities in Python and TypeScript, beyond traditional scanning.
Smarter finding prioritization: more granular coverage reporting and clearer ranking so the findings that matter most surface first.

Redesigned patch review — edit any line inline, an embedded review timeline, and a clearer diff view with a legend.
New metrics command center and refined filters in the code-security workbench.
Friendlier messages when a fix produces no changes, or when an AI usage limit is reached.
Recover a disconnected agent's identity without a full re-registration.

Resolved a cross-origin error that could cause contact and demo forms on the website to fail.

Security reports & quality gates: versioned analysis reports with automatic pass/fail release-readiness checks.
AI Insights deep links: insight findings link straight to detailed reports, with automatic CVE cross-linking.
Software bill of materials (SBOM): list and summarize dependency components, filterable by ecosystem, license, and type.
Logic Map & code graph: automatic mapping of application routes, data flows, and architecture across frameworks.
Scoped report exports: choose curated, raw, or all-inclusive data when exporting.
CyfroCompliance drill-down: move from frameworks to individual controls and findings with supporting evidence.
Inactive-user lifecycle: restore or anonymize disabled users and resend invitations.

Refreshed add-ons marketplace with a new hero, category filtering, and cleaner cards.
Clearer notifications — customer-facing labels, richer context such as the agent or scan name, and a restyled alert experience. AI notifications are now "CyfroAI Insights".
Richer dashboard previews for code security and compliance.
Disabled scans can no longer be run by mistake, with clearer status indicators.
Subscription management with grouped feature matrices and plan status.

Plans & feature entitlements: granular control over which capabilities (AI modes, scanners, billing visibility, compliance) each plan includes.
Scan targeting: run scans against specific branches or commits.
Notification center & broadcasts: in-app notifications with admin broadcast messaging.
SIEM ingestion & search: a dedicated module for log ingestion and search.
Guided signup: a multi-step onboarding wizard for your profile, organization, and team.
Sales & demo intake: website contact and demo booking with automatic email confirmations.

Refreshed admin console with a new component library, sidebar, and command palette.
Usage limits grouped into clear, categorized views.
AI cost shown as a usage percentage rather than raw figures.
Password-first signup with a direct password-setup step.
Data-preserving agent deletion, so historical data stays intact.
New platform-health monitoring and dashboards.