The Cyber Toolchain — issue 145, November 20, 2026
THE CYBER TOOLCHAIN
NO. 145
Tail
THE DAILY RELEASE FIREHOSE
PUBLISHED NOVEMBER 20, 2026 · EVERY DAY
EDITIONStail|grep|head|ls|diff|uniq
The daily firehose — everything the toolchain shipped today, already filtered.
We read every release from the 1326 tools on our watchlist at the source — GitHub and GitLab release notes, vendor release pages and changelogs, project blogs and feeds, vendor press releases, and the source code behind the tag. Bug-fix-only releases and non-product newsroom noise are dropped; what's left is summarized down to the new capability, how to try it, and any screenshots or videos the release itself published. Every entry links to the sources it was built from.
READ AS
Focus:What features shipped and what problems they solve, plus launches, packaging and where the market is moving.
Leaves out:Command-line flags, API-only deltas, usage examples, and releases that are only refinements.
Focus:New security controls, new APIs, and product improvements you can put to work, plus anything an upgrade breaks.
Leaves out:Launch and marketing news, governance, pricing, and documentation changes with no operational effect.
Focus:Dashboards, reporting, and anything you can quantify: metrics, coverage, risk scores, compliance impact, vendor moves.
Leaves out:Settings, flags, API and command-line detail, library bumps, and minor releases with no reporting or posture angle.
Focus:API, SDK and MCP changes, new libraries, breaking changes in what you import, and cyber tooling.
Leaves out:Governance, SOC operations, and vendor launches with no developer surface.
AMOUNT
DETAIL
ORDER
This issue came out before editions were ranked, so it shows the Standard edition in issue order.
What stands out across today's releases, grouped by what it lets you do. Every tool named links to its entry below.
Three unrelated things stand out. Contrast Security shipped a runtime block for Java4Shell (CVE-2026-114938) that stops the exploit at Base64 decoding, buying teams cover until the JDK patch lands. GitLab now triages and remediates vulnerabilities across a whole group and auto-rotates leaked tokens. MIND can block outbound email at Exchange Online, closing the gap left by endpoint and browser DLPDLPData Loss Prevention: a category of controls that inspect and block sensitive data from leaving an organisation via email, web, or removable media, giving cyber tools a policy enforcement layer over data in motion and at rest..
protect
Block an unpatched Java exploit at runtime instead of waiting on the JDK
Contrast's Java agent now has a block rule for Java4Shell (CVE-2026-114938) that intervenes at the Base64 decoding stage, which is the only lever most teams have while a JDK fix is still upstream and emergency redeploys are off the table. Trend Vision One adds FIPSFIPSA set of US federal standards published by NIST that define approved cryptographic algorithms and modules; cyber tools use it to meet compliance requirements in government and regulated environments. mode control on the Service Gateway appliance via CLISHCLISHA shell-like CLI framework used in network and security appliances to expose a restricted, menu-driven command interface to operators, hiding the underlying OS while enforcing role-based command access. and enforces VPCVPCA logically isolated virtual network inside a public cloud provider's infrastructure, letting tools deploy resources with private addressing, controlled routing, and firewall rules without managing physical hardware. placement for File Security's AWS Lambda functions, so scanning workloads stop having a default route to the internet.
Hand group-wide vulnerability triage and token rotation to an agent
GitLab's wave moves triage and remediation from per-project to group scope and rotates leaked tokens automatically — the window between a secret hitting a commit and someone noticing is where most credential incidents live. Endor Labs adds endorctl commands for scanning and remediating container base images, which is where a large share of inherited CVEs actually sit, plus SBOMSBOMA software bill of materials, an inventory of the components and versions inside an application. It helps teams find affected software when a dependency has a vulnerability or licensing issue./VEXVEXA machine-readable statement from a vendor saying whether a CVE in a component actually affects their product. It is how a scanner is told to stop flagging a vulnerability that is not reachable. evidence for FedRAMPFedRAMPA US government program that standardizes cloud security authorization, letting vendors get a single assessment accepted across federal agencies instead of repeating the process for each one..
Stop sensitive data on the way out of the mail server, not just the endpoint
Real-time outbound DLP blocking for Microsoft Exchange Online adds email as an enforcement point alongside endpoint and browser. Email has always been the leakiest channel and the hardest to cover from the endpoint alone, since webmail and mobile clients bypass agent-based controls entirely.
Query OpenShift compliance posture in plain English instead of reading Compliance Operator output
Elastic polls the OpenShiftOpenShiftRed Hat's Kubernetes distribution that adds developer workflows, built-in CI/CD, and stricter security defaults on top of upstream Kubernetes, giving cyber tools a managed container platform with enterprise auth and RBAC already wired in. Compliance Operator APIs into Elasticsearch, builds latest-state indices and dashboards, and puts an Agent Builder agent on top — compliance scan results stop being a point-in-time artifact someone exports for an auditor. Sublime scopes custom roles to specific message verdicts within each flag category, so a junior analyst can work malicious mail without seeing everything the detection stack flagged.
Three unrelated things stand out today: Contrast Security can block Java4Shell exploitation before a JDK patch, GitLab pushes vulnerability triage and remediation across groups, and MIND adds real-time outbound email DLPDLPData Loss Prevention: a category of controls that inspect and block sensitive data from leaving an organisation via email, web, or removable media, giving cyber tools a policy enforcement layer over data in motion and at rest. blocking for Exchange Online.
protect
Block a live exploit while the patch path catches up
Contrast added a Java agent rule for CVE-2026-114938 that stops the Java4Shell exploit at the Base64 decoding stage. That gives teams a runtime control to reduce exposure before they can safely roll a JDK patch everywhere.
Move vulnerability cleanup from finding to fix planning
GitLab added group-wide vulnerability triage and remediation, Endor Labs added endorctl container base-image remediation plus SBOMSBOMA software bill of materials, an inventory of the components and versions inside an application. It helps teams find affected software when a dependency has a vulnerability or licensing issue./VEXVEXA machine-readable statement from a vendor saying whether a CVE in a component actually affects their product. It is how a scanner is told to stop flagging a vulnerability that is not reachable. evidence, and Hacktron added configurable pull-request remediation for findings. The common win is less handoff between scanner output and engineering work: findings can be routed toward fixes, evidence, or PR workflow instead of sitting in a report.
MIND added real-time outbound email DLP blocking for Microsoft Exchange Online. Teams can now enforce policy at send time for email, rather than discovering the violation after delivery or relying only on endpoint and browser controls.
Turn compliance posture into queryable operational data
Elastic Security Labs now pulls OpenShiftOpenShiftRed Hat's Kubernetes distribution that adds developer workflows, built-in CI/CD, and stricter security defaults on top of upstream Kubernetes, giving cyber tools a managed container platform with enterprise auth and RBAC already wired in. Compliance Operator results into Elasticsearch with latest-state indices, dashboards, and a plain-English investigation agent. LogicGate added audited GRCGRCGovernance, Risk, and Compliance — a framework for aligning IT controls with business policy, regulatory requirements, and risk management so organizations can audit and report on their posture systematically. agents with configurable triggers across risk and continuity workflows, cutting some of the manual collection and follow-up work that usually surrounds governance evidence.
Hacktron adds a signed Webhook notifier for automations, a configurable pull-request remediation workflow for findings, multi-application repository membership, and a cluster of refinements to Whitebox cost estimates, scan details, PDF reports and findings-list controls.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Hacktron is a review, Whitebox, the API, and MCP.
Hacktron adds a signed Webhook notifier for automations, a configurable pull-request remediation workflow for findings, multi-application repository membership, and a cluster of refinements to Whitebox cost estimates, scan details, PDF reports and findings-list controls.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ WHAT SHIPPED ·
9 FEATURES0 OF 9 FEATURESmost completely described firstwhat's the number?
Each feature carries 0–100 for how completely the vendor documented
it — not how big or important the work is. A major capability
described in eight words scores low, and that is the finding.
depth0–40what it does and how it works or what changed
specificity0–30names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point
Code and config rank first by construction: a documented endpoint,
flag or config key scores at the top of specificity and actionability,
so it sorts above a feature described only in prose. Under
60 the notes go thin — everything below the “thinner
coverage” line is thinner documentation, not smaller work.
Hover any meter for that feature's three sub-scores.
0100
Webhook notifier for automationsNot checked yetnew85how completely this was documenteddepth30/40specificity25/30actionability30/3085 / 100
Automations gain a Webhook notifier that posts signed JSON events to an HTTPSHTTPSHTTP carried over an encrypted TLS connection. It protects traffic in transit and lets a browser verify the server's certificate. endpoint you configure, with deliveries recorded in run history; new-vulnerability automations can now notify without requiring an attached action.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Route new-vulnerability alerts from an automation directly to your incident-response pipeline without attaching a remediation action — useful for read-only notification workflows.
📍In the automation editor, open the Settings tab, add a Webhook notifier, paste your HTTPS endpoint URL, and save. Under trigger, select New vulnerability — no action is required. Deliveries will appear in run history.
— Named mechanism (signed JSON, HTTPS endpoint) with exact setup stepssnapshot-20261120
0200
Remediate workflow with PR or patch defaultNot checked yetnew80how completely this was documenteddepth30/40specificity25/30actionability25/3080 / 100
The findings list gets a Fix column showing View fix when a fix is in review or Remediate on hover for open findings without one. Clicking Remediate lets an organisation admin choose Create a pull request or Prepare the fix only and save that choice as the organisation default for future runs.
Let org admins standardise whether engineers get a ready-to-merge pull request or a local patch when they click Remediate on a finding.
📍As an organisation admin, open a finding and click Remediate, choose Create a pull request or Prepare the fix only, then save as the organisation default for future runs.
— Exact UI labels and admin steps to set the defaultsnapshot-20261120
thinner coverage below
0300
Delete finished Whitebox scansNot checked yetnew55how completely this was documenteddepth20/40specificity15/30actionability20/3055 / 100
Finished Whitebox scans and their reports can now be deleted via Delete scan in the scans list; the scan's findings are retained in Findings.
Inactive findings stay visible with undoNot checked yetimproved55how completely this was documenteddepth25/40specificity15/30actionability15/3055 / 100
Findings moved to an inactive status now remain visible but dimmed in the findings list, with an Undo option shown in a Moved to inactive toast; the findings list no longer applies a default Status filter.
Whitebox cost estimates by code sizeNot checked yetimproved50how completely this was documenteddepth25/40specificity15/30actionability10/3050 / 100
Whitebox cost estimates now quote the whole pentest by code size, showing each repository's lines of code with an expandable folder breakdown of Lines and Credits.
Multi-application repository membershipNot checked yetnew45how completely this was documenteddepth20/40specificity15/30actionability10/3045 / 100
Repositories can now belong to several applications at once; repository pickers, including Add context, display every application a repository belongs to, with a +N badge on hover.
RedAmon is an open-source, self-hosted AI penetration testing framework: maps your attack surface into a graph, autonomously exploits it from a Kali sandbox with human approval gates, and opens PRs that fix what it finds. MCP both ways: plug in any MCP server as a tool, or drive RedAmon from Claude Code or your own agent.
RedAmon's v7.9.0 release adds a project-aware AI Navigator chat assistant and overhauls Tradecraft into a 2.0 per-account knowledge library with three new agent tools, replacing 1.x in a breaking migration, while also shipping MCPMCPModel Context Protocol, a standard for connecting AI applications to external tools and data sources. A compatible server exposes capabilities an assistant can use through a common interface. tool-plugin presets, a stricter Neo4jNeo4jA graph database maintained by Neo4j Inc. that stores data as nodes and relationships rather than tables, letting cyber tools query complex entity connections and traversal paths efficiently.APOCAPOCA query language and stored-procedure system for Neo4j graph databases, letting cyber tools run complex graph traversals and algorithms that plain Cypher cannot express. procedure set, and expanded accounts/roles documentation.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ WHAT SHIPPED ·
5 FEATURES1 OF 5 FEATURESmost completely described firstwhat's the number?
Each feature carries 0–100 for how completely the vendor documented
it — not how big or important the work is. A major capability
described in eight words scores low, and that is the finding.
depth0–40what it does and how it works or what changed
specificity0–30names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point
Code and config rank first by construction: a documented endpoint,
flag or config key scores at the top of specificity and actionability,
so it sorts above a feature described only in prose. Under
60 the notes go thin — everything below the “thinner
coverage” line is thinner documentation, not smaller work.
Hover any meter for that feature's three sub-scores.
0101
Tradecraft 2.0 knowledge library and agent toolsNot checked yetbreaking98how completely this was documenteddepth38/40specificity30/30actionability30/3098 / 100
RedAmon introduces Tradecraft 2.0, a per-account knowledge library of trusted offensive-security sites the agent searches and reads mid-attack chain, managed under Global Settings → Tradecraft.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄ Sites can be added via 105 presets or custom URL ingestion through Check URL, using Live (site-map and live page reads) or Stored (cleaned page copies in your database) ingestion modes, with site logins and a Map view to browse, search, pin, block, rename, and restore pages; Stored pages count against a quota of 1 GB per account and 20 GB per install, with 95 configurable ingestion and search settings. Three new agent tools query the library: tradecraft_search (keyword search with no model call), tradecraft_lookup (reads one section of a page by its page ID, e.g. patt:412), and tradecraft_ask (a librarian returning up to three checked pages and exact excerpts for a vague need). It also supports GitHub-token-authenticated CVE archive indexing (up to 20,000 largest CVE files, no text stored), docs-site mapping from search indexes or sitemaps, rule-based, LLMLLMA large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code.-guided, or Auto website crawls, and a focus mode (keep everything, Focused on chosen topics, or Tag only), with resources enabled per project via the Agent Skills & Tradecraft tab. This replaces Tradecraft 1.x for every account with no option to keep 1.x — resources convert automatically on first agent start after update; the old page-cache volume redamon_tradecraft_cache is deleted on update (run touch .tradecraft-cache-removed at the repository root beforehand to keep it); rolling back to 7.8.0 drops the Tradecraft 2.0 tables and data; projects that previously limited tradecraft_lookup to exploitation and post-exploitation phases now have it expanded to all three, logged as system.tradecraft_phases_backfill; and new projects start with the CVE (MSF) skill off by default.
Add HackTricks as a Tradecraft source so the agent can read the exact technique page it needs mid-attack chain.
📍Go to Global Settings → Tradecraft, click Check URL, enter https://book.hacktricks.xyz, choose Stored ingestion mode, and save — the site is ingested in the background and the agent will surface relevant pages as Library now during attack steps.
Add a Stored Tradecraft resource and restrict it to specific topics so the agent draws only relevant pages during an engagement.
📍Go to Global Settings → Tradecraft, add a new resource, set its type to Stored, set the ingestion focus to Focused, and select the topics to keep. Then open the project's Agent Skills & Tradecraft tab and enable the resource for that project.
— Names every config key, tool, quota and migration step with exact identifiers.v7.9.0
0200
Project-aware AI Navigator chat assistantNot checked yetnew80how completely this was documenteddepth30/40specificity25/30actionability25/3080 / 100
AI Navigator is a chat assistant available on every page that answers questions about how RedAmon works (linking to documentation) and about the selected project's data — including the graph, findings, scans, settings, triage, remediations, reports, and workspace files. Access is controlled via the new navigator:use permission, enabled or restricted instance-wide under IAMIAMIdentity and Access Management — a framework of policies and services for controlling who can authenticate to systems and what resources they are authorized to use, letting cyber tools enforce least-privilege and audit access. → Instance and granted per project through the share dialog; each user picks their own AI model at Global Settings → Models by feature → AI Navigator. Custom permission sets show navigator:use as available, not granted, so admins must manually add it to sets that should use Navigator, and rolling back to 7.8.0 drops every Navigator conversation along with every owner's collaborator choice.
Grant a collaborator access to AI Navigator on your project so they can query project findings and graph data via the chat.
📍In the project share dialog, find the collaborator's entry and enable the Navigator toggle (requires the navigator:use permission to be present in their role). To allow or block Navigator for the entire instance, go to IAM → Instance and toggle the Navigator setting.
Pick the AI model that powers AI Navigator for your account so the chat can answer questions about your project data.
📍Go to Global Settings → Models by feature → AI Navigator and select your preferred model. If no model is set, the first Navigator message will prompt you to choose one.
During a pentest, ask AI Navigator what findings are open in the current project without leaving your working context.
📍In any RedAmon page, open the AI Navigator chat panel and type: 'What critical findings are in this project?' — it will answer using only the data you are permitted to read, citing the exact documentation section and project graph.
— Exact permission and settings path given, but model list and scope limits unstated.v7.9.0
0300
MCP server tool plugins with presetsNot checked yetnew72how completely this was documenteddepth25/40specificity25/30actionability22/3072 / 100
Any Model-Context-Protocol server can be plugged into the agent as a tool plugin, with 33 prefilled presets including Shodan, GitHub, Hugging Face, Censys, and mitmproxy. Each plugin offers a Test/Discover button to verify the connection and phase-aware Tool Matrix integration so it can be enabled for specific attack phases.
Connect a Shodan MCP server so the agent can query internet-exposed assets as a tool plugin during recon phases.
📍Go to the agent's tool plugin settings, select the Shodan preset from the 33 prefilled Model-Context-Protocol presets, click Test/Discover to verify the connection, then enable it in the Tool Matrix for the Recon phase.
— Names presets and control, but no API schema or limits given.product docs
0400
Neo4j APOC procedure restrictionNot checked yetbreaking60how completely this was documenteddepth20/40specificity28/30actionability12/3060 / 100
Neo4j now loads only the APOC procedures RedAmon calls, so scripts calling APOC procedures not used by RedAmon get a no procedure registered error — apoc.load.* and apoc.cypher.* are no longer available.
— Names exact removed procedure families but gives no migration path.v7.9.0
thinner coverage below
0500
Accounts and roles documentation expandedNot checked yetimproved44how completely this was documenteddepth18/40specificity18/30actionability8/3044 / 100
Documentation now covers all 99 permissions explained step by step, including signing in, inviting people, resetting credentials, disabling and deleting accounts, managing administrators, giving permission sets, and acting as a user.
— Documentation-only update with no new mechanism or command.product docs
└──▷ BREAKING ON UPGRADE
!
Tradecraft 2.0 replaces Tradecraft 1.x for every account with no option to keep 1.x; resources are converted automatically on first agent start after update.
!
Scripts calling APOC procedures not used by RedAmon will get no procedure registered — Neo4j now loads only the APOC procedures RedAmon calls, so apoc.load.* and apoc.cypher.* are no longer available.
!
Rolling back to 7.8.0 drops every AI Navigator conversation, every owner's collaborator choice, and the Tradecraft 2.0 tables and columns with their data.
!
New projects now start with the CVE (MSF) skill off by default.
!
The old Tradecraft page-cache volume (redamon_tradecraft_cache) is deleted on update; run touch .tradecraft-cache-removed at the repository root before updating to keep it.
!
Custom permission sets show navigator:use as available, not granted — admins must manually add it to sets that should use AI Navigator.
!
Where tradecraft_lookup was allowed in exactly exploitation and post-exploitation phases, it is automatically expanded to all three phases; the audit log's system.tradecraft_phases_backfill row lists affected projects.
BloodHound added configurable limits for PostgreSQL graph traversals to control resource usage during analysis.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ WHAT SHIPPED ·
1 FEATURE0 OF 1 FEATUREmost completely described firstwhat's the number?
Each feature carries 0–100 for how completely the vendor documented
it — not how big or important the work is. A major capability
described in eight words scores low, and that is the finding.
depth0–40what it does and how it works or what changed
specificity0–30names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point
Code and config rank first by construction: a documented endpoint,
flag or config key scores at the top of specificity and actionability,
so it sorts above a feature described only in prose. Under
60 the notes go thin — everything below the “thinner
coverage” line is thinner documentation, not smaller work.
Hover any meter for that feature's three sub-scores.
0100
Configurable traversal limits for PostgreSQLNot checked yetnew50how completely this was documenteddepth20/40specificity20/30actionability10/3050 / 100
BloodHound now supports configurable limits for PostgreSQL graph traversals to control traversal state and byte usage, available via updated DAWGS v1.0.2.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
— Names DAWGS version but no config key or flag givenv9.10.1
OpenQodex is an open source AI code review for coding agents. A separate reviewer checks every change your agent writes before you push, in Claude Code, Codex, Cursor and Cline.
OpenQodex's 0.22.0 release expands its code graph with gin service support and deeper TypeScript resolution, while improving coverage-gap reporting and fixing import/reference accounting bugs across languages.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ WHAT SHIPPED ·
5 FEATURES0 OF 5 FEATURESmost completely described firstwhat's the number?
Each feature carries 0–100 for how completely the vendor documented
it — not how big or important the work is. A major capability
described in eight words scores low, and that is the finding.
depth0–40what it does and how it works or what changed
specificity0–30names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point
Code and config rank first by construction: a documented endpoint,
flag or config key scores at the top of specificity and actionability,
so it sorts above a feature described only in prose. Under
60 the notes go thin — everything below the “thinner
coverage” line is thinner documentation, not smaller work.
Hover any meter for that feature's three sub-scores.
0100
Gin service support in `graph routes`Not checked yetnew90how completely this was documenteddepth30/40specificity30/30actionability30/3090 / 100
openqodex graph routes now reads gin services, including engines, groups with their prefixes, Use middleware, and handlers; inline route handlers (e.g.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄ app.get('/x', (c) => ...)) are treated as first-class definitions, and the new --exclude-tests flag omits routes whose test files build the application.
Audit only production routes in a gin-based Go service, skipping routes defined inside test files.
$ openqodex graph routes --exclude-tests
— Names exact command and flag with runnable exampleopenqodex@0.22.0
0200
Expanded TypeScript type reference resolutionNot checked yetimproved83how completely this was documenteddepth30/40specificity28/30actionability25/3083 / 100
openqodex graph references on a TypeScript type now finds uses in interface properties, type aliases, type arguments (e.g. new Map<string, Repo>()), type parameter constraints and defaults, and classes that implement or extend the type.
Find all call sites and implementors of a TypeScript interface, including structurally-matched classes and type-argument uses that were previously missed.
$ openqodex graph references ISessionStore
— Named command, scope, and runnable example providedopenqodex@0.22.0
0300
Coverage gap reporting for unread languagesNot checked yetnew78how completely this was documenteddepth35/40specificity28/30actionability15/3078 / 100
Code graph answers now surface coverage gaps for unread languages — Rust, C, C++, C#, Kotlin, Swift, PHP, and Scala — with reasons like 2 Rust files not read; each answer gains an unknown.groups field breaking gaps down by cause with a count and up to two examples, plus a listIncomplete flag, and the CLI prints certain (list incomplete) in the counts line.
Per-import-line counting in `graph packages`Not checked yetimproved65how completely this was documenteddepth25/40specificity25/30actionability15/3065 / 100
openqodex graph packages now counts every import line separately, so a file importing the same package via import type and import no longer loses the duplicate site; importers still lists each file once.
External classification for chained TypeScript callsNot checked yetimproved55how completely this was documenteddepth25/40specificity20/30actionability10/3055 / 100
TypeScript method calls on values returned by a package are now classified as external at any chain depth, instead of raising an unknown gap that set the floor for every same-named repository method.
— Explains mechanism and prior bug, lacks concrete exampleopenqodex@0.22.0
GitHub expanded its public preview of the pull request Overview page, adding clearer merge-requirement and check signaling, force-push commit comparison links with timestamps, new sidebar actions for locking conversations and editing development links, and smaller UI and Copilot refinements.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
GitHub is a web-based platform for version control and collaborative software development using Git repositories.
GitHub expanded its public preview of the pull request Overview page, adding clearer merge-requirement and check signaling, force-push commit comparison links with timestamps, new sidebar actions for locking conversations and editing development links, and smaller UI and Copilot refinements.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ WHAT SHIPPED ·
5 FEATURES0 OF 5 FEATURESmost completely described firstwhat's the number?
Each feature carries 0–100 for how completely the vendor documented
it — not how big or important the work is. A major capability
described in eight words scores low, and that is the finding.
depth0–40what it does and how it works or what changed
specificity0–30names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point
Code and config rank first by construction: a documented endpoint,
flag or config key scores at the top of specificity and actionability,
so it sorts above a feature described only in prose. Under
60 the notes go thin — everything below the “thinner
coverage” line is thinner documentation, not smaller work.
Hover any meter for that feature's three sub-scores.
0100
Clearer merge requirements and check previewsNot checked yetimproved55how completely this was documenteddepth25/40specificity15/30actionability15/3055 / 100
In the Overview page, failed optional checks no longer make mergeable pull requests appear blocked, previously missing merge requirements now show what needs attention, and checks and approvals now expand when clicked instead of pinning a preview.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Force-push comparison links and commit timestampsNot checked yetimproved50how completely this was documenteddepth20/40specificity15/30actionability15/3050 / 100
In the Overview page, force-push activity now links to the old and new commits so you can compare changes, and commit activity now includes timestamps so you can see when changes were made.
Opt in to the Overview page public preview to access the new sidebar actions and commit comparison links on your next pull request.
📍In the pull request header, select the ✨ Try the Overview page button. To return to the classic view, use the same menu.
Sidebar actions for locking and development linksNot checked yetnew35how completely this was documenteddepth15/40specificity10/30actionability10/3035 / 100
The Overview page gained sidebar actions to lock or unlock PR conversations directly in the view, and to edit Development sidebar links without leaving the page.
Responsive branch display on Overview pageNot checked yetimproved20how completely this was documenteddepth10/40specificity5/30actionability5/3020 / 100
Source and target branches now stay visible on smaller screens in the Overview page.
GitHub adds live repository migration from Azure DevOps, syncing history, branch policies as rulesets, Boards connections, and Pipelines with continuous sync until cutover.
Migrate repositories from Azure DevOps Services to GitHub Enterprise Cloud with data residency while developers keep contributing, then cut over at a time you choose. Not checked yet
Manage migrations through the Azure DevOps portal for a guided experience or script them with the Azure DevOps CLI. Not checked yet
Pre-migration validation and lifecycle tracking identify issues and monitor progress before cutover. Not checked yet
Sync Git history, branches, tags, and pull requests — including titles, descriptions, comments, and user history — from Azure DevOps to GitHub. Not checked yet
Automatically migrate Azure DevOps branch policies as GitHub branch rulesets. Not checked yet
+1 moreshow less
Optionally create the Azure Boards connection and rewire Azure Pipelines to the migrated GitHub repository to reduce manual setup after cutover. Not checked yet
GitHub Actions adds RHELRHELRed Hat Enterprise Linux, a commercially supported Linux distribution maintained by Red Hat, chosen by cyber tools for its long support lifecycle, certified hardware compatibility, and enterprise deployment prevalence. 9 and RHEL 10 runner images as generally available for production workflows on Team and Enterprise Cloud plans.
Cve-lite-cli is a fast, developer-friendly JS/TS dependency vulnerability scanner with local lockfile scanning, OSV matching, direct vs transitive visibility, --fix, JSON output, and practical remediation guidance.
cve-lite-cli adds multi-format SBOMSBOMA software bill of materials, an inventory of the components and versions inside an application. It helps teams find affected software when a dependency has a vulnerability or licensing issue. output, SBOM advisory database metadata, --allow-empty for empty-project scans, and output sanitisation against workflow command injectioncommand injectionAn attack that makes an application execute operating-system commands supplied through untrusted input. It often happens when input is passed unsafely to a shell..
└──▷ GET THIS VERSIONnpmgit
npm
$ npm install -g cve-lite-cli@1.42.0
git
$ git clone --branch v1.42.0 https://github.com/OWASP/cve-lite-cli.git
# already have the repo? check out this version:$ git checkout v1.42.0
named in the project's README · checked 20 Nov
Request multiple SBOM formats in one run by passing --sbom cyclonedx,spdx or repeating the flag (--cdx --sbom spdx); each format is written to a separate file sharing one timestamp. Not checked yet
CycloneDXCycloneDXAn open standard for describing software components and their dependencies, vulnerabilities and other supply-chain information. Tools use it to exchange bills of materials and related evidence. and SPDXSPDXAn open standard for describing software components, licenses and their relationships. Its documents let different tools exchange a software bill of materials without inventing a new format.SBOMsSBOMsA software bill of materials, an inventory of the components and versions inside an application. It helps teams find affected software when a dependency has a vulnerability or licensing issue. now record advisory database metadata: whether the scan ran offline, when the local database was last synced, and where it came from. Not checked yet
Allow empty-project scans without an exit code change by passing --allow-empty or setting the GitHub Action's allow-empty input. Not checked yet
!Node 22.13.0 or later (22.x line) or 23.4.0 or later is now required; earlier Node versions are no longer supported.
!A scan that finds no lockfile and no exact-pinned package.json now exits 3 instead of 0, with or without --fail-on; pass --allow-empty or set the Action's allow-empty input to restore the old behaviour.
!cve-lite --report <dir> with no project path is now refused; use --report=<dir> or put the project path before --report.
!Some CycloneDX severity ratings come out lower because a severity no longer carries over from one advisory to another for the same package.
!The first scan after upgrading runs with a cold OSVOSVOpen Source Vulnerabilities, a vulnerability database and schema maintained by Google that gives cyber tools a machine-readable, cross-ecosystem format for querying and sharing CVE-style advisories. cache because the cache format changed.
!A known malicious package (MAL- advisory) now fails --fail-on at any level, including the default, where it previously passed every threshold except low.
Contrast Security is a secure AI-generated code with Contrast. Our runtime application security platform observes real behavior to block attacks in real time.
Contrast Security shipped a runtime block rule for the Java4Shell exploit (CVE-2026-114938), letting the Java agent stop attacks at the Base64 decoding stage ahead of a JDK patch.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ WHAT SHIPPED ·
1 FEATURE1 OF 1 FEATUREmost completely described firstwhat's the number?
Each feature carries 0–100 for how completely the vendor documented
it — not how big or important the work is. A major capability
described in eight words scores low, and that is the finding.
depth0–40what it does and how it works or what changed
specificity0–30names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point
Code and config rank first by construction: a documented endpoint,
flag or config key scores at the top of specificity and actionability,
so it sorts above a feature described only in prose. Under
60 the notes go thin — everything below the “thinner
coverage” line is thinner documentation, not smaller work.
Hover any meter for that feature's three sub-scores.
0101
Java4Shell runtime block rule for Java agentNot checked yetnew86how completely this was documenteddepth30/40specificity28/30actionability28/3086 / 100
Contrast Java agent 6.38.0 or later adds a Java4Shell rule that can block or monitor CVE-2026-114938 exploit attempts at the Base64 decoding stage inside the JVMJVMThe Java Virtual Machine, a runtime maintained by Oracle (and implemented by others) that executes Java bytecode on any OS, giving cyber tools a portable, garbage-collected environment with a large ecosystem of libraries., stopping attacks before patching is complete.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄ As an interim workaround on hosts where neither patching nor agent deployment is possible, the JVM flag -XX:-UseBASE64Intrinsics can be set (e.g. java -XX:-UseBASE64Intrinsics -jar your-service.jar), at a possible performance cost.
Put internet-facing Java services into block mode immediately while waiting for JDK vendors to ship fixed builds.
Qodana-cli's v2026.3.1 release adds an air-gapped scanning mode for internet-isolated environments and expands the Poly image's language coverage with Lua support and auto-detection.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Qodana-cli jetbrains Qodana’s official command line tool
Qodana-cli's v2026.3.1 release adds an air-gapped scanning mode for internet-isolated environments and expands the Poly image's language coverage with Lua support and auto-detection.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ WHAT SHIPPED ·
2 FEATURES0 OF 2 FEATURESmost completely described firstwhat's the number?
Each feature carries 0–100 for how completely the vendor documented
it — not how big or important the work is. A major capability
described in eight words scores low, and that is the finding.
depth0–40what it does and how it works or what changed
specificity0–30names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point
Code and config rank first by construction: a documented endpoint,
flag or config key scores at the top of specificity and actionability,
so it sorts above a feature described only in prose. Under
60 the notes go thin — everything below the “thinner
coverage” line is thinner documentation, not smaller work.
Hover any meter for that feature's three sub-scores.
0100
Air-gapped scanning modeNot checked yetnew85how completely this was documenteddepth30/40specificity25/30actionability30/3085 / 100
Enable air-gapped mode for environments with no internet access using the --air-gapped flag on qodana scan, or persist it project-wide via the airGapped field in qodana.yaml so every CI run scans offline automatically.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Scan a project in an air-gapped environment where the analysis container has no outbound internet access.
$ qodana scan --air-gapped
Persist air-gapped mode in your project config so every scan on any CI runs offline automatically.
yaml
airGapped: true
— Names exact flag and config field with runnable examplesv2026.3.1
thinner coverage below
0200
Lua and Dart support in Poly imageNot checked yetnew45how completely this was documenteddepth20/40specificity15/30actionability10/3045 / 100
Adds the Lua plugin to the Qodana Poly image, extending its language coverage, and adds auto-detection of Lua and Dart projects so the CLI now suggests the Qodana Poly linter when these are the dominant languages.
— Describes what changed but no command or config shownv2026.3.1
Deepsec is a security harness for finding vulnerabilities in your codebase powered by coding agents
deepsec adds chunked sandbox tarball uploads and now retries Codex turns that end without a completed result︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Uploads sandbox tarballs larger than 32 MiB in parts via uploadTarballToSandbox, writing each chunk as <remoteTarPath>.part-NNNNN and joining them with cat on the sandbox, so scans of large repositories no longer fail with Request Entity Too Large errors.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄ Not checked yet
Codex batches whose turn ends without turn.completed — for example when OpenAI's safety filter blocks a turn mid-way — now retry and then fail instead of recording incomplete agent messages as clean findings. Not checked yet
Uploads sandbox tarballs larger than 32 MiB in parts via uploadTarballToSandbox, writing each chunk as <remoteTarPath>.part-NNNNN and joining them with cat on the sandbox, so scans of large repositories no longer fail with Request Entity Too Large errors.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄ Not checked yet
Teams and agents ship trusted software faster with GitLab, the intelligent orchestration platform with context and governance for the entire software lifecycle.
GitLab's November 19 wave is almost entirely about agentic engineering: group-wide vulnerability triage and remediation, a Slack-native agent, scheduled flow automations, a new Business Logic Security Analyzer, automatic leaked-token rotation, and a set of coding-agent skills for MR follow-up, code review, and issue-to-MR delivery.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ WHAT SHIPPED ·
14 FEATURES2 OF 14 FEATURESmost completely described firstwhat's the number?
Each feature carries 0–100 for how completely the vendor documented
it — not how big or important the work is. A major capability
described in eight words scores low, and that is the finding.
depth0–40what it does and how it works or what changed
specificity0–30names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point
Code and config rank first by construction: a documented endpoint,
flag or config key scores at the top of specificity and actionability,
so it sorts above a feature described only in prose. Under
60 the notes go thin — everything below the “thinner
coverage” line is thinner documentation, not smaller work.
Hover any meter for that feature's three sub-scores.
0101
Leaked PAT auto-rotation via Secrets ManagerNot checked yetnew95how completely this was documenteddepth35/40specificity30/30actionability30/3095 / 100
Enable the rotation workflow on GitLab Self-Managed with the secrets_manager_pat_rotation feature flag (on by default on GitLab.com, not yet on GitLab Dedicated).︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄ The leaked-token-rotation-automation tool reads GitLab PATsPATsA personal access token, a credential used by scripts and integrations instead of a user's password. Its permissions and expiry depend on the service issuing it. — including fine-grained PATs — from CI/CD project variables, which secret detection does not scan, rotates them via the token's own api or self_rotate scope, stores the new values in Secrets Manager, and opens an MR for any jobs that reference the variable; cloud provider keys and other secret types are not yet covered.
Execute batch rotation across a group, capping at 25 rotations in one run to stay within credits budget.
Duo setup CLI skills and security triage-and-fix promptNot checked yetnew90how completely this was documenteddepth30/40specificity30/30actionability30/3090 / 100
The gitlab-duo-setup CLI now installs the gitlab-mr-followup skill, which checks open review threads, failed jobs, and conflicts on a merge request and works through them from your agent session, with both skills living in ~/.claude/skills/ or .claude/skills/ inside the repo for team sharing and MCPMCPModel Context Protocol, a standard for connecting AI applications to external tools and data sources. A compatible server exposes capabilities an assistant can use through a common interface. connectivity verifiable via /mcp and the list_duo_agents_and_flows tool call; on GitLab Ultimate, gitlab-duo-setup also offers an optional security findings prompt that, when answered yes, enables the security triage-and-fix workflow and adds security flows to the group.
Bootstrap a top-level group for agentic engineering — installs the setup and MR-followup skills, then runs the guided setup from inside Claude Code.
Bootstrap the full Duo agent workflow for your team so every MR gets automated review and pipeline repair by default.
📍Install the skills from https://gitlab.com/gitlab-da/projects/agentic-eng-config and run gitlab-duo-setup for my GitLab group. Show me the plan before you change anything.
Agentic @GitLab in SlackNot checked yetnew85how completely this was documenteddepth35/40specificity25/30actionability25/3085 / 100
Mention @GitLab in a Slack channel or DM to have an agent read the conversation, create issues and merge requests, and start agent flows, with results posted back to the thread; ask it to summarise a channel with @GitLab, summarize #channel for me. Each request runs as the requester's linked GitLab account, scoped to their existing permissions, creates an auditable agent session, and draws from GitLab Credits. It does not operate in external Slack Connect channels.
Turn a live Slack channel discussion into a tracked GitLab issue without leaving the thread.
📍Steps: 1. Open the Slack channel where the discussion is happening. 2. Type '@GitLab create an issue from this thread' as a message. 3. The agent reads the conversation, creates the issue in GitLab under your account permissions, and posts a link back to the thread.
Catch up on a busy channel before asking the agent to act, so the resulting work item reflects the full discussion.
📍Steps: 1. In any public or private Slack channel, type @GitLab, summarize #channel for me. 2. Review the summary the agent posts back. 3. Follow up with '@GitLab create a merge request for the agreed approach' to convert the decision into tracked work.
Scheduled automations for Duo Agent Platform flowsNot checked yetnew85how completely this was documenteddepth35/40specificity25/30actionability25/3085 / 100
Schedule Duo Agent Platform flows to run recurring tasks — from every 15 minutes to every month, at a chosen time and timezone — with no event trigger required (beta in GitLab 19.6). The redesigned Automations page (previously called Triggers) shows run volume, success rate, and session details for both scheduled and event-based automations. Each scheduled automation runs as a dedicated service account with scoped permissions, so agent actions appear separately from human actions in audit events, and every change to an automation is itself recorded in audit events.
Turn on experimental and beta features and set up a first scheduled automation — for example, an hourly CI health investigation — to verify the beta is active on your instance.
📍In GitLab 19.6, go to your project › Automations (formerly Triggers) › New automation. Describe the task in plain language (e.g. 'Check open CI incidents and comment with the likely root cause'), set the cadence to Every hour, choose your timezone, and save. Confirm the automation appears with run count and success rate on the Automations page.
Group-wide ATR profiles for vulnerability triage and remediationNot checked yetnew80how completely this was documenteddepth30/40specificity25/30actionability25/3080 / 100
Apply automated triage and remediation (ATR) profiles across many projects at once from Security Inventory or the security configuration page, selecting by group or attribute, without scripting against the API. Profiles offer three presets — Conservative, Standard, or Proactive — covering false-positive detection, risk contextualization, and vulnerability resolution.
Roll out agentic triage and remediation to every project in a group without writing API scripts — pick a preset that matches your risk appetite.
📍In GitLab Ultimate, go to Security › Security Inventory (or the security configuration page), select projects by group or attribute, choose an ATR profile preset (Conservative, Standard, or Proactive), and click Apply.
Vulnerability Context Flow exploitability risk tableNot checked yetnew75how completely this was documenteddepth30/40specificity25/30actionability20/3075 / 100
Enable Vulnerability Context Flow to add a risk table to SASTSASTStatic application security testing, analyzing source code or compiled code for vulnerabilities without running the application. It helps developers find problems before deployment. findings showing whether exploiting a flaw requires authentication or elevated authorization, whether the service handles sensitive data, and whether it is exposed to untrusted input, so teams can rank the backlog by exploitability as well as severity.
Add exploitability context to SAST findings so your team can prioritize public-facing services over isolated internal ones without manual triage.
📍In GitLab Ultimate 19.6, follow the Vulnerability Context Flow setup to enable the feature; once on, open any SAST finding to see the new risk table showing authentication requirements, sensitive data exposure, and untrusted-input exposure.
Business Logic Security Analyzer (BLSA)Not checked yetnew65how completely this was documenteddepth25/40specificity25/30actionability15/3065 / 100
Self-service beta scans repositories and merge requests for broken ownership logic, missing authorization checks, multi-step abuse paths, and other authorization/access-control vulnerabilities that pattern-based scanners miss. Available via self-service on GitLab.com (GitLab Ultimate); Self-Managed and Dedicated customers must contact GitLab for access, and it can be powered by Anthropic's Claude Mythos 5.2 on request.
Catch broken ownership logic and missing authorization checks that static scanners miss by enabling Business Logic Security Analyzer on GitLab.com without waiting for an account team.
📍In GitLab Ultimate on GitLab.com, follow the business logic scanning setup to self-enable BLSA; for Self-Managed or Dedicated, contact GitLab for access.
GitLab Credits tracking, caps, and efficiency improvementsNot checked yetimproved60how completely this was documenteddepth25/40specificity20/30actionability15/3060 / 100
Prompt caching on the Duo Agent Platform lets eligible tasks use fewer model requests, stretching GitLab Credits up to 2x further; individual GitLab Credits consumption can now be tracked on GitLab.com filtered by product and date range without needing an administrator report; the Duo Agent Platform Impact dashboard (introduced in GitLab 19.5) shows adoption, work activity, and credit consumption broken down by model, flow type, and group or project; and a default per-user credit cap with individual overrides can be set on the Credit caps page (introduced in GitLab 19.4 for GitLab.com and Self-Managed) without using an API.
/goal objective planning in Duo CLI and ChatNot checked yetnew60how completely this was documenteddepth25/40specificity15/30actionability20/3060 / 100
Describe an objective with /goal in the Duo CLI or Duo Agent Platform Chat to have GitLab Duo plan, execute, and judge completion of the work, with the ability to pause, change, or redirect the agent at any point.
Hand off a multi-step objective to Duo so it plans, executes, and self-judges completion while you stay in control.
$ /goal Triage all open issues labelled 'needs-triage', assign severity, and add a comment summarising the action taken on each.
Developer flow: issue to merge requestNot checked yetnew60how completely this was documenteddepth25/40specificity15/30actionability20/3060 / 100
Hand an assigned issue to the Developer flow and Duo runs in GitLab, reads the issue and repository, writes the change, and opens an MR in the user's name; governance and approval rules can still enforce a human review.
Hand a well-scoped issue to Duo's Developer flow so it writes the code and opens an MR without you copying context into a prompt.
📍Find my assigned issues in <project>. Hand the best-scoped one to the Duo Developer flow.
Review Duo's pipeline fix and code review feedback on a specific MR before deciding what to merge.
📍Did Duo propose a fix for the failed pipeline on MR !42? Show me the diff and why it should fix the failure.
Automated $0.25 second-look code reviewNot checked yetnew50how completely this was documenteddepth20/40specificity20/30actionability10/3050 / 100
Duo now gives every eligible merge request an automated second-look code review for a flat $0.25 per review when the MR is marked ready, posting feedback on the diff.
Endor Labs shipped new endorctl commands for scanning and remediating container base images alongside SBOMSBOMA software bill of materials, an inventory of the components and versions inside an application. It helps teams find affected software when a dependency has a vulnerability or licensing issue./VEXVEXA machine-readable statement from a vendor saying whether a CVE in a component actually affects their product. It is how a scanner is told to stop flagging a vulnerability that is not reachable. evidence generation for FedRAMPFedRAMPA US government program that standardizes cloud security authorization, letting vendors get a single assessment accepted across federal agencies instead of repeating the process for each one. compliance, and launched a deep integration with OpenAI's Daybreak Defense Factory that feeds reachability-based remediation guidance into Codex Security.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Endor Labs is an application security platform built for AI-generated and human-written code, helping teams prioritize and fix real risk across the SDLC.
Endor Labs shipped new endorctl commands for scanning and remediating container base images alongside SBOMSBOMA software bill of materials, an inventory of the components and versions inside an application. It helps teams find affected software when a dependency has a vulnerability or licensing issue./VEXVEXA machine-readable statement from a vendor saying whether a CVE in a component actually affects their product. It is how a scanner is told to stop flagging a vulnerability that is not reachable. evidence generation for FedRAMPFedRAMPA US government program that standardizes cloud security authorization, letting vendors get a single assessment accepted across federal agencies instead of repeating the process for each one. compliance, and launched a deep integration with OpenAI's Daybreak Defense Factory that feeds reachability-based remediation guidance into Codex Security.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ WHAT SHIPPED ·
3 FEATURES3 OF 3 FEATURESmost completely described firstwhat's the number?
Each feature carries 0–100 for how completely the vendor documented
it — not how big or important the work is. A major capability
described in eight words scores low, and that is the finding.
depth0–40what it does and how it works or what changed
specificity0–30names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point
Code and config rank first by construction: a documented endpoint,
flag or config key scores at the top of specificity and actionability,
so it sorts above a feature described only in prose. Under
60 the notes go thin — everything below the “thinner
coverage” line is thinner documentation, not smaller work.
Hover any meter for that feature's three sub-scores.
0101
Base image update scanning and remediation commandsNot checked yetnew90how completely this was documenteddepth30/40specificity30/30actionability30/3090 / 100
Scan a container and check for available base image updates with endorctl container scan --base-image-check-updates (plus --namespace, --image, and --project-name), then list candidate base image updates with endorctl container remediation list-base-image-updates (same arguments), which shows findings fixed and introduced by severity and reachability for each option.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Check a container image for available base image updates before shipping, to see which upgrade resolves the most reachable findings without introducing new ones.
Review the ranked list of candidate base image updates — with per-severity and per-reachability counts of findings fixed and introduced — to pick the safest upgrade path.
OpenAI Daybreak Defense Factory integration with Codex SecurityNot checked yetnew80how completely this was documenteddepth35/40specificity30/30actionability15/3080 / 100
An AI SASTSASTStatic application security testing, analyzing source code or compiled code for vulnerabilities without running the application. It helps developers find problems before deployment. agent gives Codex evidence and remediation guidance for vulnerabilities in first-party source code; an SCASCASoftware composition analysis, identifying third-party dependencies and checking them for known vulnerabilities and license issues. It covers software a team uses rather than writes itself. Remediation agent gives Codex reachability evidence and remediation guidance for open source vulnerabilities, cutting findings by up to 99% via full-stack reachability across direct and transitive dependencies; container base image vulnerabilities also get Codex reachability evidence and remediation guidance. A code context graph tells agents how the application uses each dependency, reducing agent token usage by 92% and completing security tasks 19x faster in Endor Labs benchmarks, and real-time vulnerability intelligence screens more than 50,000 open source packages per month for malicious behavior (including risks without a CVE), surfacing this data to Codex. Setup follows the OpenAI Daybreak Defense Factory import guide: open the console, select Endor Labs in the integrations section, authenticate, and findings with reachability, risk, and remediation guidance appear in Codex for automated fix implementation.
Follow the Daybreak Defense Factory import guide to connect Endor Labs findings to Codex Security for end-to-end automated remediation.
📍Steps: 1. Open the OpenAI Daybreak Defense Factory console. 2. Navigate to the integrations section and select Endor Labs. 3. Follow the import guide to authenticate and configure the connection. 4. Endor Labs findings — with reachability, risk, and remediation guidance — will appear in Codex for automated fix implementation.
SBOM/VEX evidence and finding deduplication for complianceNot checked yetnew40how completely this was documenteddepth20/40specificity15/30actionability5/3040 / 100
Endor Labs now deduplicates findings across SCA and container scans and generates SBOM and VEX evidence attached to every vulnerability verdict, producing machine-readable reporting intended for FedRAMP VDRVDRA Virtual Data Room is a secure, hosted document repository used for controlled sharing of sensitive files during due diligence, audits, or investigations, giving cyber tools a governed environment to exchange evidence or reports./VER rules.
OpenBao's latest releases add per-mount control over entity/group alias case-sensitivity and extend the Transit key-export API to support external-key public keys.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
OpenBao is a software solution to manage, store, and distribute sensitive data including secrets, certificates, and keys.
OpenBao's latest releases add per-mount control over entity/group alias case-sensitivity and extend the Transit key-export API to support external-key public keys.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ WHAT SHIPPED ·
2 FEATURES1 OF 2 FEATURESmost completely described firstwhat's the number?
Each feature carries 0–100 for how completely the vendor documented
it — not how big or important the work is. A major capability
described in eight words scores low, and that is the finding.
depth0–40what it does and how it works or what changed
specificity0–30names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point
Code and config rank first by construction: a documented endpoint,
flag or config key scores at the top of specificity and actionability,
so it sorts above a feature described only in prose. Under
60 the notes go thin — everything below the “thinner
coverage” line is thinner documentation, not smaller work.
Hover any meter for that feature's three sub-scores.
0101
Per-mount alias case-sensitivity controlNot checked yetbreaking75how completely this was documenteddepth30/40specificity25/30actionability20/3075 / 100
Auth mounts can now control alias case-sensitivity using the case_sensitive_alias config key.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄ Entity and group aliases on newly created auth mounts are now case-sensitive by default, whereas previously they were case-insensitive; existing mounts are unaffected and can use case_sensitive_alias to override the new default.
— Names exact config key and before/after behaviour changev2.7.3v2.6.6
thinner coverage below
0200
Public key export for Transit external-key typeNot checked yetimproved40how completely this was documenteddepth15/40specificity15/30actionability10/3040 / 100
The existing key-export API can now export public keys for Transit keys of type external-key.
— Names API and key type but no endpoint path or usage detailv2.7.3
└──▷ BREAKING ON UPGRADE
!
Entity and group aliases on newly created auth mounts are now case-sensitive by default, which may break setups that relied on case-insensitive alias matching.
!
Entity and group aliases of newly created auth mounts are now case-sensitive by default (previously case-insensitive). Existing mounts are unaffected; use case_sensitive_alias to override.
TruffleHog adds a Baseten API key detector with permission analysis, surfaces upsert errors from AWS Secrets Manager sources, and adds new metrics for detectors that exceed their timeout.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
TruffleHog finds, verifies, and analyzes leaked credentials across Git, chat, wikis, cloud storage, and more
TruffleHog adds a Baseten API key detector with permission analysis, surfaces upsert errors from AWS Secrets Manager sources, and adds new metrics for detectors that exceed their timeout.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ WHAT SHIPPED ·
3 FEATURES0 OF 3 FEATURESmost completely described firstwhat's the number?
Each feature carries 0–100 for how completely the vendor documented
it — not how big or important the work is. A major capability
described in eight words scores low, and that is the finding.
depth0–40what it does and how it works or what changed
specificity0–30names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point
Code and config rank first by construction: a documented endpoint,
flag or config key scores at the top of specificity and actionability,
so it sorts above a feature described only in prose. Under
60 the notes go thin — everything below the “thinner
coverage” line is thinner documentation, not smaller work.
Hover any meter for that feature's three sub-scores.
0100
Baseten API key detector with permission analysisNot checked yetnew81how completely this was documenteddepth25/40specificity28/30actionability28/3081 / 100
Adds a Baseten API key detector, enabled via feature.BasetenDetectorEnabled, with permission analysis available through baseten.AnalyzeAndPrintPermissions using the baseten case in the analyzer CLI.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
— Names exact flag, function, and CLI casev3.100.0
0200
Detector timeout overrun metricsNot checked yetnew70how completely this was documenteddepth20/40specificity30/30actionability20/3070 / 100
Adds metrics for detectors that run past their timeout: a trufflehog_scanner_detector_timeout_ignored_total counter, a trufflehog_scanner_detector_workers_over_timeout gauge, and an overrun-duration histogram, all labelled by detector_name.
— Names exact metric names and labels, no usage examplev3.100.0
thinner coverage below
0300
Upsert error field in AWS Secrets Manager sourceNot checked yetimproved50how completely this was documenteddepth15/40specificity20/30actionability15/3050 / 100
Adds upsert_error to AWSSecretsManager source metadata fields, surfacing upsert failures in scan output.
— Names exact field but minimal mechanism detailv3.100.0
WPScan v4.2.1 adds visibility into responses truncated by --max-response-size, flagging incomplete scan data instead of silently missing it.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
WPScan WordPress security scanner. Written for security professionals and blog maintainers to test the security of their WordPress websites. Contact us via contact@wpscan.com
WPScan v4.2.1 adds visibility into responses truncated by --max-response-size, flagging incomplete scan data instead of silently missing it.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ WHAT SHIPPED ·
1 FEATURE0 OF 1 FEATUREmost completely described firstwhat's the number?
Each feature carries 0–100 for how completely the vendor documented
it — not how big or important the work is. A major capability
described in eight words scores low, and that is the finding.
depth0–40what it does and how it works or what changed
specificity0–30names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point
Code and config rank first by construction: a documented endpoint,
flag or config key scores at the top of specificity and actionability,
so it sorts above a feature described only in prose. Under
60 the notes go thin — everything below the “thinner
coverage” line is thinner documentation, not smaller work.
Hover any meter for that feature's three sub-scores.
0100
Truncated response reportingNot checked yetnew83how completely this was documenteddepth30/40specificity28/30actionability25/3083 / 100
Responses cut short by --max-response-size are now listed in the scan summary: a count warning appears in the CLI output (with full URLs shown when --verbose is passed), and a truncated_responses array is added to JSON output, so findings based on incomplete bodies are flagged rather than silently missed.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
— Names exact flags and JSON field; mechanism and purpose both explained.v4.2.1
Wallarm rolled out seven AI Agentic Operators that automate alert triage, rule hygiene, and policy tuning, manageable via its MCPMCPModel Context Protocol, a standard for connecting AI applications to external tools and data sources. A compatible server exposes capabilities an assistant can use through a common interface. server and with an optional autonomous execution mode.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Discover the latest features, improvements, and updates in Wallarm API Security
Wallarm rolled out seven AI Agentic Operators that automate alert triage, rule hygiene, and policy tuning, manageable via its MCPMCPModel Context Protocol, a standard for connecting AI applications to external tools and data sources. A compatible server exposes capabilities an assistant can use through a common interface. server and with an optional autonomous execution mode.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ WHAT SHIPPED ·
3 FEATURES0 OF 3 FEATURESmost completely described firstwhat's the number?
Each feature carries 0–100 for how completely the vendor documented
it — not how big or important the work is. A major capability
described in eight words scores low, and that is the finding.
depth0–40what it does and how it works or what changed
specificity0–30names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point
Code and config rank first by construction: a documented endpoint,
flag or config key scores at the top of specificity and actionability,
so it sorts above a feature described only in prose. Under
60 the notes go thin — everything below the “thinner
coverage” line is thinner documentation, not smaller work.
Hover any meter for that feature's three sub-scores.
0100
Seven AI Agentic Operators for triage and tuningNot checked yetnew65how completely this was documenteddepth25/40specificity30/30actionability10/3065 / 100
Wallarm introduces seven Agentic Operators: Radar, which detects attack spikes, investigates them, and suggests a response; Sift, which clears false positives on an hourly cycle so alerts remain actionable; Trim, which tidies the active ruleset to prevent rule sprawl; Tune, which adjusts protection policies to match observed application behaviour; Lens, which reviews the API inventory and flags risky endpoints; Bard, which converts the current security posture into a human-readable briefing; and Wally, which answers questions and handles configuration requests on demand.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
MCP server control of Agentic OperatorsNot checked yetnew55how completely this was documenteddepth20/40specificity15/30actionability20/3055 / 100
All Agentic Operators can be managed through Wallarm's MCP server, letting them be plugged into external agentic workflows (triggering operators like Sift, Radar, or Wally) without returning to the Wallarm console.
Plug Wallarm Agentic Operators into an external agentic workflow via the MCP server so your own agents can trigger Sift, Radar, or Wally without opening the Wallarm console.
📍Contact your customer success team to enable Agentic Operators, then connect your agentic workflow to Wallarm's MCP server to invoke operators such as Sift, Radar, and Wally programmatically.
Autonomous mode with audit loggingNot checked yetnew35how completely this was documenteddepth20/40specificity10/30actionability5/3035 / 100
Agentic Operators support an autonomous mode in which routine, approved actions run without manual confirmation, with every finding logged to the audit trail.
Fastly is a cloud edge computing platform that provides CDN, DDoS protection, load balancing, and real-time content delivery for websites and applications.
Fastly's Next-Gen WAFWAFA reverse proxy that inspects HTTP/HTTPS traffic and blocks requests matching known attack patterns such as SQL injection or XSS, sitting between clients and a web application to filter malicious input before it reaches the server. agent 4.84.0 improves Envoy/extproc observability and strengthens auth-related threat detection.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ WHAT SHIPPED ·
2 FEATURES0 OF 2 FEATURESmost completely described firstwhat's the number?
Each feature carries 0–100 for how completely the vendor documented
it — not how big or important the work is. A major capability
described in eight words scores low, and that is the finding.
depth0–40what it does and how it works or what changed
specificity0–30names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point
Code and config rank first by construction: a documented endpoint,
flag or config key scores at the top of specificity and actionability,
so it sorts above a feature described only in prose. Under
60 the notes go thin — everything below the “thinner
coverage” line is thinner documentation, not smaller work.
Hover any meter for that feature's three sub-scores.
0100
Envoy/extproc connection metrics and version reportingNot checked yetimproved40how completely this was documenteddepth15/40specificity15/30actionability10/3040 / 100
The Next-Gen WAF agent's Envoy/extproc module now adds connection metrics and Envoy version reporting.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Sublime Security is an email security platform that detects and blocks phishing, malware, and other email-based threats using customizable, human-readable detection rules.
Sublime Security added verdict-based RBACRBACRole-based access control, where permissions are assigned to roles such as reader or administrator and users receive those roles. It makes access easier to manage than individual permission lists., letting custom roles be scoped to specific message verdicts within each flag category.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ WHAT SHIPPED ·
1 FEATURE0 OF 1 FEATUREmost completely described firstwhat's the number?
Each feature carries 0–100 for how completely the vendor documented
it — not how big or important the work is. A major capability
described in eight words scores low, and that is the finding.
depth0–40what it does and how it works or what changed
specificity0–30names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point
Code and config rank first by construction: a documented endpoint,
flag or config key scores at the top of specificity and actionability,
so it sorts above a feature described only in prose. Under
60 the notes go thin — everything below the “thinner
coverage” line is thinner documentation, not smaller work.
Hover any meter for that feature's three sub-scores.
0100
Verdict-based role access controlNot checked yetnew78how completely this was documenteddepth30/40specificity28/30actionability20/3078 / 100
Custom roles can now restrict which message verdicts — Malicious, Suspicious, Unknown, Error, Graymail, Spam, or Benign — a role can view, configured per message flag type (flagged, unflagged, user reports, DLPDLPData Loss Prevention: a category of controls that inspect and block sensitive data from leaving an organisation via email, web, or removable media, giving cyber tools a policy enforcement layer over data in motion and at rest. flagged) when creating or editing a custom role under Settings > Roles.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Limit a tier-1 analyst role so it can only open Spam and Graymail messages — keeping Malicious and Suspicious content restricted to senior responders
📍In the platform, go to Settings › Roles, create or edit a custom role, and under each message flag type (Flagged, Unflagged, User Reports, DLP Flagged) select only the verdicts that role should see (e.g. Spam, Graymail) — leaving Malicious, Suspicious, Unknown, Error, and Benign unchecked.
Plerion's November changelog adds six new end-of-life checks spanning AWS, GCP and Kubernetes services, tightens auto-triage of likely-exploited vulnerabilities, and fixes false-pass behavior in public-accessibility checks for RDS and EMR.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Plerion is a cloud security platform that provides unified visibility, risk prioritization, and remediation across cloud environments to help organizations manage and reduce security risks.
Plerion's November changelog adds six new end-of-life checks spanning AWS, GCP and Kubernetes services, tightens auto-triage of likely-exploited vulnerabilities, and fixes false-pass behavior in public-accessibility checks for RDS and EMR.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ WHAT SHIPPED ·
2 FEATURES0 OF 2 FEATURESmost completely described firstwhat's the number?
Each feature carries 0–100 for how completely the vendor documented
it — not how big or important the work is. A major capability
described in eight words scores low, and that is the finding.
depth0–40what it does and how it works or what changed
specificity0–30names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point
Code and config rank first by construction: a documented endpoint,
flag or config key scores at the top of specificity and actionability,
so it sorts above a feature described only in prose. Under
60 the notes go thin — everything below the “thinner
coverage” line is thinner documentation, not smaller work.
Hover any meter for that feature's three sub-scores.
0100
End-of-life checks for cloud data and compute servicesNot checked yetnew80how completely this was documenteddepth30/40specificity30/30actionability20/3080 / 100
Six new checks flag services running past end of support: PLERION-AWS-1264 for EMR clusters on a release past end of standard support (covering release labels such as emr-6.15.0 and emr-spark-8.2.0 and AMI versions), PLERION-GCP-354 for Dataproc clusters on an image version past end of support (judged by major/minor image version), PLERION-K8S-187 for Kubernetes nodes whose kubelet minor version is past upstream end of life (skipping vendor builds like EKSEKSAmazon's managed Kubernetes service, run by AWS, that handles control-plane provisioning so teams can deploy containerized workloads without operating the underlying cluster infrastructure themselves., GKEGKEGoogle Kubernetes Engine, Google's managed Kubernetes service, which lets cyber tools run containerized workloads without managing the underlying cluster infrastructure., k3s, RKE2, OpenShiftOpenShiftRed Hat's Kubernetes distribution that adds developer workflows, built-in CI/CD, and stricter security defaults on top of upstream Kubernetes, giving cyber tools a managed container platform with enterprise auth and RBAC already wired in., AKSAKSAzure Kubernetes Service, Microsoft's managed Kubernetes offering, which lets cyber tools offload cluster provisioning and scaling to Azure so engineers focus on workload logic instead of control-plane operations.), PLERION-GCP-355 for Cloud Composer environments on an unsupported image version (covering Composer 1, 2, and 3), PLERION-AWS-1265 for Neptune clusters on an engine version past end of life (covering four-part versions such as 1.2.1.0 and 1.4.6.1), and PLERION-AWS-1266 for CodeBuild projects using a deprecated build image (including standard:3.0, amazonlinux2-x86_64-standard:2.0, standard:4.0, amazonlinux2-aarch64-standard:1.0, amazonlinux2-x86_64-standard:3.0, and all WINDOWS_CONTAINER environment types).︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄ All mark failed findings as End of life.
EPSS-based severity floor in auto-triageNot checked yetimproved70how completely this was documenteddepth30/40specificity25/30actionability15/3070 / 100
The auto-triage rules Memory safety, no known exploit and Kernel, no known exploit now leave a vulnerability at its advisory severity when its EPSS score is above 66%, reversing any prior reduction to Medium.
eksctl v0.232.0 adds a flag to skip CloudFormationCloudFormationAn AWS service that lets you describe and provision cloud infrastructure as declarative YAML or JSON templates, so a cyber tool can reproducibly spin up or tear down its entire AWS environment with a single API call. termination protection and a new cluster config field for elastic load balancing in Kubernetes network settings, alongside tighter IRSAIRSAAn AWS mechanism that lets a Kubernetes pod assume an IAM role via a service account token, so cyber tools running in EKS can call AWS APIs without storing long-lived credentials. scoping for CSICSIAn ambiguous abbreviation — it most commonly means Container Storage Interface (a Kubernetes storage plugin standard) or Crime Scene Investigation — without more context the correct meaning cannot be determined reliably. add-ons.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Eksctl is the official CLI for Amazon EKS
eksctl v0.232.0 adds a flag to skip CloudFormationCloudFormationAn AWS service that lets you describe and provision cloud infrastructure as declarative YAML or JSON templates, so a cyber tool can reproducibly spin up or tear down its entire AWS environment with a single API call. termination protection and a new cluster config field for elastic load balancing in Kubernetes network settings, alongside tighter IRSAIRSAAn AWS mechanism that lets a Kubernetes pod assume an IAM role via a service account token, so cyber tools running in EKS can call AWS APIs without storing long-lived credentials. scoping for CSICSIAn ambiguous abbreviation — it most commonly means Container Storage Interface (a Kubernetes storage plugin standard) or Crime Scene Investigation — without more context the correct meaning cannot be determined reliably. add-ons.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ WHAT SHIPPED ·
4 FEATURES0 OF 4 FEATURESmost completely described firstwhat's the number?
Each feature carries 0–100 for how completely the vendor documented
it — not how big or important the work is. A major capability
described in eight words scores low, and that is the finding.
depth0–40what it does and how it works or what changed
specificity0–30names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point
Code and config rank first by construction: a documented endpoint,
flag or config key scores at the top of specificity and actionability,
so it sorts above a feature described only in prose. Under
60 the notes go thin — everything below the “thinner
coverage” line is thinner documentation, not smaller work.
Hover any meter for that feature's three sub-scores.
0100
Elastic load balancing in Kubernetes network configNot checked yetnew75how completely this was documenteddepth20/40specificity25/30actionability30/3075 / 100
A new elasticLoadBalancing field under kubernetesNetworkConfig lets you configure elastic load balancing in cluster network settings, e.g.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄ setting enabled: true in cluster.yaml.
Enable elastic load balancing settings in your cluster's Kubernetes network configuration.
— Named config field with runnable YAML examplev0.232.0
0200
CloudFormation termination protection opt-out flagNot checked yetnew70how completely this was documenteddepth15/40specificity25/30actionability30/3070 / 100
The new --cfn-disable-termination-protection flag lets you opt out of CloudFormation stack termination protection.
— Names exact flag but no further mechanism detailv0.232.0
thinner coverage below
0300
Default EKS version when unsetNot checked yetimproved50how completely this was documenteddepth20/40specificity15/30actionability15/3050 / 100
When --version is not set, eksctl now uses the EKS default Kubernetes version instead of requiring an explicit value.
— Explains behaviour change but no command examplev0.232.0
0400
Tighter IRSA scoping for CSI and EMR rolesNot checked yetimproved50how completely this was documenteddepth20/40specificity20/30actionability10/3050 / 100
IRSA trust for EBSEBSAmazon's block-storage service for EC2 instances, providing persistent network-attached volumes that survive instance restarts and can be snapshotted, giving cyber tools durable, scalable disk without managing physical drives. and EFS CSI add-on roles now scopes to the add-on service account, reducing over-permissioning, and the EMR containers role now includes get and list secrets permissions.
— Describes permission change with named roles, no exact policy shownv0.232.0
Databricks is rolling out OAuth 2.0OAuth 2.0An open authorization framework (IETF RFC 6749) that lets a service grant limited API access to a third party without sharing credentials, giving cyber tools a standard way to authenticate against external platforms. authentication for webhook notification destinations in compliance-profile workspaces and has made ai_forecast version 2 generally available with holiday effects, external covariates, and non-negative constraints.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Databricks is a unified data analytics platform that enables collaborative data engineering, machine learning, and analytics using Apache Spark in the cloud.
Databricks is rolling out OAuth 2.0OAuth 2.0An open authorization framework (IETF RFC 6749) that lets a service grant limited API access to a third party without sharing credentials, giving cyber tools a standard way to authenticate against external platforms. authentication for webhook notification destinations in compliance-profile workspaces and has made ai_forecast version 2 generally available with holiday effects, external covariates, and non-negative constraints.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ WHAT SHIPPED ·
2 FEATURES0 OF 2 FEATURESmost completely described firstwhat's the number?
Each feature carries 0–100 for how completely the vendor documented
it — not how big or important the work is. A major capability
described in eight words scores low, and that is the finding.
depth0–40what it does and how it works or what changed
specificity0–30names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point
Code and config rank first by construction: a documented endpoint,
flag or config key scores at the top of specificity and actionability,
so it sorts above a feature described only in prose. Under
60 the notes go thin — everything below the “thinner
coverage” line is thinner documentation, not smaller work.
Hover any meter for that feature's three sub-scores.
0100
ai_forecast version 2 generally availableNot checked yetnew62how completely this was documenteddepth25/40specificity22/30actionability15/3062 / 100
ai_forecast version 2 in Databricks SQL is now generally available, adding built-in holiday effects, external covariates, and non-negative forecast constraints, powered by a research-optimized time series foundation model for improved out-of-the-box accuracy.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
OAuth authentication for webhook destinationsNot checked yetUpdated integrationnew53how completely this was documenteddepth25/40specificity18/30actionability10/3053 / 100
In workspaces with the compliance security profile enabled, Databricks will soon enable OAuth 2OAuth 2An open authorization framework (RFC 6749) that lets a service grant limited access to a user's resources on another service without exposing credentials, giving cyber tools a standard way to authenticate against third-party APIs..0 client-credential authentication for webhook destinations by default — Databricks fetches an access tokenaccess tokenA credential that authorizes a client to use a service or API. It usually has limited permissions and a lifetime; anyone holding it may be able to use those permissions. from a configured OAuth token endpoint and attaches it to each notification request.
Elastic Security Labs is a powered by Elasticsearch. Ready to deploy your data?
Elastic Security Labs shipped a new OpenShiftOpenShiftRed Hat's Kubernetes distribution that adds developer workflows, built-in CI/CD, and stricter security defaults on top of upstream Kubernetes, giving cyber tools a managed container platform with enterprise auth and RBAC already wired in. compliance integration that polls the OpenShift Compliance Operator APIs into Elasticsearch, builds latest-state indices and dashboards from that data, and layers an Elastic Agent Builder agent on top for plain-English posture investigation.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ WHAT SHIPPED ·
5 FEATURES0 OF 5 FEATURESmost completely described firstwhat's the number?
Each feature carries 0–100 for how completely the vendor documented
it — not how big or important the work is. A major capability
described in eight words scores low, and that is the finding.
depth0–40what it does and how it works or what changed
specificity0–30names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point
Code and config rank first by construction: a documented endpoint,
flag or config key scores at the top of specificity and actionability,
so it sorts above a feature described only in prose. Under
60 the notes go thin — everything below the “thinner
coverage” line is thinner documentation, not smaller work.
Hover any meter for that feature's three sub-scores.
0100
OpenShift compliance data ingestion via HTTP JSON inputNot checked yetnew90how completely this was documenteddepth35/40specificity30/30actionability25/3090 / 100
Polls OpenShift Compliance Operator APIs every 60 seconds via the HTTP JSON input, shipping compliancescans.compliance.openshift.io, compliancecheckresults.compliance.openshift.io, and complianceremediations.compliance.openshift.io resources into Elasticsearch data streams.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄ Every ingested event is tagged with orchestrator.cluster.name using the CLUSTER_NAME environment variable set on the Elastic Agent, keeping multi-cluster scan results separated in Elasticsearch.
Deploy Elastic Agent on an OpenShift cluster to start collecting compliance scan data into Elasticsearch, tagging every event with the cluster name so multi-cluster results stay separated.
Upload the custom OpenShift Security Posture Management integration package through the Kibana Fleet UI to register all data streams, ingest pipelines, transforms, and dashboards in one step.
📍In Kibana, go to Fleet › Integrations › Upload Integration, select the OSPM zip package from the GitHub repository, then open Fleet › Agents, create a new Agent Policy, add the OSPM integration, set the three compliance endpoints and 60-second polling interval, generate an enrollment token, and deploy the manifest to your cluster.
Latest-state transforms for compliance indicesNot checked yetnew68how completely this was documenteddepth30/40specificity28/30actionability10/3068 / 100
Three Elasticsearch transforms — scans, checks, and remedy — continuously collapse time-series compliance data into latest-state indices named ospm-compliance_-latest, deduplicating check events by cluster and metadata.name.
Prebuilt Kibana dashboards for OpenShift complianceNot checked yetnew60how completely this was documenteddepth20/40specificity25/30actionability15/3060 / 100
Adds three prebuilt Kibana dashboards — [OpenShift Compliance] Posture Overview, [OpenShift Compliance] Overall Results, and [OpenShift Compliance] Auto Remediations — installed automatically when the integration package is uploaded through Fleet.
CEL custom rules for manual compliance checksNot checked yetnew50how completely this was documenteddepth20/40specificity20/30actionability10/3050 / 100
Manual compliance rules can be evaluated with Common Expression Language (CELCELCommon Expression Language, an open-source expression language maintained by Google, used to embed lightweight, safe policy evaluation logic into tools without running a full scripting engine.) custom rules, with a scheduled Elastic Workflow annotating results back to the manual rule in the cluster.
Elastic Agent Builder agent for compliance investigationNot checked yetnew35how completely this was documenteddepth15/40specificity10/30actionability10/3035 / 100
Adds an Elastic Agent Builder agent over the centralised compliance data so teams can query fleet-wide posture and investigate control failures in plain English.
Tines changed how Cases exports are delivered, moving from email attachments to direct in-app downloads.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Tines is an apps, agents and automations, in the complete environment for enterprise AI.
Tines changed how Cases exports are delivered, moving from email attachments to direct in-app downloads.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ WHAT SHIPPED ·
1 FEATURE0 OF 1 FEATUREmost completely described firstwhat's the number?
Each feature carries 0–100 for how completely the vendor documented
it — not how big or important the work is. A major capability
described in eight words scores low, and that is the finding.
depth0–40what it does and how it works or what changed
specificity0–30names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point
Code and config rank first by construction: a documented endpoint,
flag or config key scores at the top of specificity and actionability,
so it sorts above a feature described only in prose. Under
60 the notes go thin — everything below the “thinner
coverage” line is thinner documentation, not smaller work.
Hover any meter for that feature's three sub-scores.
0100
In-app download for Cases exportsNot checked yetbreaking40how completely this was documenteddepth15/40specificity10/30actionability15/3040 / 100
Cases exports are now downloaded directly inside Tines instead of being sent as email attachments.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄ This replaces the previous delivery method entirely, so exported files must now be retrieved from within the app.
Yeti 2.13.0 adds settings for audit-log body control and URL-import safety, tightens observable type validation, and fixes API key revocation and expiry enforcement.
└──▷ GET THIS VERSION
$ git clone --branch 2.13.0 https://github.com/yeti-platform/yeti.git
# already have the repo? check out this version:$ git checkout 2.13.0
Control whether request bodies appear in the console and audit log file with the audit_log_bodies setting or YETI_SYSTEM_AUDIT_LOG_BODIES environment variable (default: true). Not checked yet
Restrict URL imports to public addresses only by setting allow_internal_url_imports = false or YETI_SYSTEM_ALLOW_INTERNAL_URL_IMPORTS; /observables/import/url now times out at 5 s connect / 30 s between bytes and returns 400 on failure. Not checked yet
Control the number of web workers in webserver-prod with the YETI_WEB_WORKERS environment variable (default: 4). Not checked yet
Observable values submitted to POST /observables/ and POST /observables/bulk are now validated against their declared type; mismatched values are refused with a 400 (single) or skipped with a log entry (bulk). Not checked yet
/graph/match now lists an entity under every looked-up value it is linked to, returning one [relationship, entity] pair per link in entities; clients should group by the relationship's source or target field instead of de-duplicating by entity id. Not checked yet
+1 moreshow less
Event task acts_on patterns now match correctly on Python 3.11+, so all three shipped event tasks fire as expected. Not checked yet
!abuse.ch feeds (MalwareBazaar, ThreatFox, URLhaus) now require YETI_ABUSECH_AUTH_KEY or [abusech] in yeti.conf; without a key all three feeds fail.
!API key expiry logic is inverted on upgrade: keys with a future expiry date start working; keys whose expiry date has passed stop working.
!Deleted or disabled API keys sent as bearer tokensbearer tokensA credential that grants access to whoever presents it, without a separate proof of possession. Protecting the token matters because a stolen copy can be used by someone else. are now refused (previously they kept working).
!POST /observables/ and POST /observables/bulk now reject values that do not match their declared type (400 for single add; skipped with a log entry in bulk); feeds or scripts that sent mistyped values will start logging skipped items.
!/graph/matchentities now contains one [relationship, entity] pair per link instead of one per entity per request; clients that de-duplicate by entity id must switch to grouping by the relationship's source or target field.
!Event task acts_on patterns now fire on Python 3.11+, including all three shipped event tasks; DatadogMetrics will start reporting a yeti.object series with type:acl.
!webserver-prod now starts 4 workers by default (YETI_WEB_WORKERS) instead of one per host core.
!/observables/import/url now times out (5 s connect, 30 s between bytes) and returns 400 instead of 500 on a failed fetch.
OpenCTI 7.261119.0 ships a next-generation playbook editor, multi-threat-actor comparison on the MITRE ATT&CKMITRE ATT&CKA publicly maintained knowledge base of adversary tactics and techniques observed in real attacks, maintained by MITRE. Cyber tools use it as a common taxonomy to classify, map, and communicate threat behavior. matrix, dashboard-wide global filters, and a new server-level HTTP security config block, alongside configuration import/export for RBACRBACRole-based access control, where permissions are assigned to roles such as reader or administrator and users receive those roles. It makes access easier to manage than individual permission lists. and email templates.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
OpenCTI is an open Cyber Threat Intelligence Platform
OpenCTI 7.261119.0 ships a next-generation playbook editor, multi-threat-actor comparison on the MITRE ATT&CKMITRE ATT&CKA publicly maintained knowledge base of adversary tactics and techniques observed in real attacks, maintained by MITRE. Cyber tools use it as a common taxonomy to classify, map, and communicate threat behavior. matrix, dashboard-wide global filters, and a new server-level HTTP security config block, alongside configuration import/export for RBACRBACRole-based access control, where permissions are assigned to roles such as reader or administrator and users receive those roles. It makes access easier to manage than individual permission lists. and email templates.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ WHAT SHIPPED ·
7 FEATURES0 OF 7 FEATURESmost completely described firstwhat's the number?
Each feature carries 0–100 for how completely the vendor documented
it — not how big or important the work is. A major capability
described in eight words scores low, and that is the finding.
depth0–40what it does and how it works or what changed
specificity0–30names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point
Code and config rank first by construction: a documented endpoint,
flag or config key scores at the top of specificity and actionability,
so it sorts above a feature described only in prose. Under
60 the notes go thin — everything below the “thinner
coverage” line is thinner documentation, not smaller work.
Hover any meter for that feature's three sub-scores.
0100
RBAC and email template configuration presetsNot checked yetnew65how completely this was documenteddepth25/40specificity25/30actionability15/3065 / 100
RBAC configuration (groups, roles, policies) can now be exported and imported through a new Import Wizard with partial-import support, and the same configuration-preset mechanism now also covers export/import of email templates, including a new built-in User locked email template for account lock-out notifications.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
— Names RBAC fields, Import Wizard, and the new email template7.261119.0
0200
Next-generation playbook editorNot checked yetimproved60how completely this was documenteddepth30/40specificity15/30actionability15/3060 / 100
The playbook editor gains deterministic layout, node rewiring, an undo button, and a new node-deletion behaviour that re-attaches the downstream branch instead of removing it.
— Describes concrete editor mechanics but no exact UI path or command7.261119.0
thinner coverage below
0300
Server-level HTTP security config blockNot checked yetnew55how completely this was documenteddepth25/40specificity20/30actionability10/3055 / 100
A new global HTTP switch controls CSPCSPContent Security Policy: a browser security mechanism, standardized by the W3C, that lets a web server declare which sources of scripts, styles, and other resources a page may load, blocking injection attacks., rate limiting, body size, and redirects via a new server-level config block.
— Names the controls but not the exact config key7.261119.0
0400
Global filter for dashboard widgetsNot checked yetnew55how completely this was documenteddepth20/40specificity15/30actionability20/3055 / 100
A global filter can be applied across an entire dashboard to scope all widgets at once, for example by TLP marking or author, set from the dashboard editor's Filters panel.
Scope all dashboard widgets to a single knowledge domain without editing each one individually.
📍In the dashboard editor, open the Filters panel and set a global filter (for example, by TLP marking or author); all widgets will inherit the constraint automatically.
— Includes a concrete navigation path and example7.261119.0
0500
Multi-actor MITRE ATT&CK comparisonNot checked yetnew45how completely this was documenteddepth15/40specificity15/30actionability15/3045 / 100
The TTPTTPTactics, techniques and procedures, the goals, methods and specific behaviors used by an attacker. They describe how an attack works, rather than only the files or addresses it leaves behind. Analyser now lets users compare multiple threat actors side-by-side on the MITREMITREA US federally funded research and development center that maintains public frameworks—ATT&CK, CVE, D3FEND—that cyber tools use to classify threats, vulnerabilities, and defensive techniques in a common language.ATT&CKATT&CKA publicly maintained knowledge base by MITRE cataloguing adversary tactics, techniques, and procedures observed in real attacks, giving cyber tools a common taxonomy for classifying and correlating threat behavior. matrix.
— Names the feature and location, no deeper mechanism given7.261119.0
0600
EPSS score filtering with null supportNot checked yetimproved45how completely this was documenteddepth15/40specificity15/30actionability15/3045 / 100
Entities can be filtered by EPSSEPSSExploit Prediction Scoring System, an estimate of the probability that a published vulnerability will be exploited in the next 30 days. It helps prioritize investigation and patching. score, and results where the score is null are now included.
— Clear behaviour change, no config detail beyond that7.261119.0
0700
Inline image preview in entity ContentNot checked yetnew30how completely this was documenteddepth10/40specificity10/30actionability10/3030 / 100
PNG and JPEG images can now be previewed inline in the Content section of entities.
— Bare description with no further mechanism given7.261119.0
OSSEC HIDS 4.5.0 expands rule matching with new agent, location, hostname and CDB field syntax, adds a post-baseline FIMFIMFile Integrity Monitoring — a technique that tracks cryptographic hashes or metadata of files on disk and alerts when they change unexpectedly, giving cyber tools a way to detect tampering or unauthorized modification. create-tracking mode, and ships a PCRE2PCRE2A widely used open-source regular expression library maintained by Philip Hazel, offering Perl-compatible pattern matching; cyber tools use it for fast, expressive text and binary pattern searches. pattern-testing CLI alongside Windows installer flags, daily log rotation, IISIISMicrosoft's web server for Windows, bundled with the OS and managed via the Windows service infrastructure; cyber tools target or integrate with it because it hosts a large share of enterprise web applications. log field extraction, an OpenSSLOpenSSLAn open-source cryptography library and toolkit, maintained by the OpenSSL Project, that provides TLS, SSL, and general-purpose crypto primitives so tools avoid reimplementing certificate handling and cipher negotiation from scratch. 4 build of ossec-authd, and broader notify_time handling.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
OSSEC HIDS is an OSSEC is an Open Source Host-based Intrusion Detection System that performs log analysis, file integrity checking, policy monitoring, rootkit detection, real-time alerting and active response.
OSSEC HIDS 4.5.0 expands rule matching with new agent, location, hostname and CDB field syntax, adds a post-baseline FIMFIMFile Integrity Monitoring — a technique that tracks cryptographic hashes or metadata of files on disk and alerts when they change unexpectedly, giving cyber tools a way to detect tampering or unauthorized modification. create-tracking mode, and ships a PCRE2PCRE2A widely used open-source regular expression library maintained by Philip Hazel, offering Perl-compatible pattern matching; cyber tools use it for fast, expressive text and binary pattern searches. pattern-testing CLI alongside Windows installer flags, daily log rotation, IISIISMicrosoft's web server for Windows, bundled with the OS and managed via the Windows service infrastructure; cyber tools target or integrate with it because it hosts a large share of enterprise web applications. log field extraction, an OpenSSLOpenSSLAn open-source cryptography library and toolkit, maintained by the OpenSSL Project, that provides TLS, SSL, and general-purpose crypto primitives so tools avoid reimplementing certificate handling and cipher negotiation from scratch. 4 build of ossec-authd, and broader notify_time handling.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ WHAT SHIPPED ·
8 FEATURES0 OF 8 FEATURESmost completely described firstwhat's the number?
Each feature carries 0–100 for how completely the vendor documented
it — not how big or important the work is. A major capability
described in eight words scores low, and that is the finding.
depth0–40what it does and how it works or what changed
specificity0–30names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point
Code and config rank first by construction: a documented endpoint,
flag or config key scores at the top of specificity and actionability,
so it sorts above a feature described only in prose. Under
60 the notes go thin — everything below the “thinner
coverage” line is thinner documentation, not smaller work.
Hover any meter for that feature's three sub-scores.
0100
Post-baseline FIM create trackingNot checked yetnew85how completely this was documenteddepth30/40specificity25/30actionability30/3085 / 100
Enable new post-baseline FIM create tracking with <report_new_files_as_changed>, which stores new files created after the initial baseline as !+++ entries visible in syscheck_control -i.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Track files created after the initial FIM baseline as changes so they appear in syscheck_control reports, rather than being silently ignored.
— Names config key, entry marker and inspection command, with runnable config.4.5.0
0200
Expanded rule-matching syntax for agents, paths and fieldsNot checked yetnew78how completely this was documenteddepth25/40specificity28/30actionability25/3078 / 100
Adds new rule-matching capabilities: expose the OSSEC agent identity as agent_name or <agent_name>, match log paths with <location> or <location_pcre2>, preserve device hostname from agent-forwarded syslog lines for <hostname> matching, and allow CDB <list> matching on decoded system_name via field="system_name".
Write rules that alert on specific OSSEC agents by name, or scope a rule to a particular log path, without relying on IP address alone.
— Names four new rule-matching syntax elements with a worked example.4.5.0
0300
PCRE2 rule pattern testing CLINot checked yetnew70how completely this was documenteddepth20/40specificity20/30actionability30/3070 / 100
Test a regex pattern against PCRE2 — with the same flags as a <pcre2> rule — using ossec-regex -p.
Verify that a PCRE2 pattern behaves as it would inside a rule before deploying it, catching flag or syntax mismatches early.
$ ossec-regex -p '^Failed password for (invalid user )?\S+ from (?P<srcip>\S+)'
— Exact runnable command given to test patterns before deploying rules.4.5.0
0400
Daily rotation of ossec.logNot checked yetimproved65how completely this was documenteddepth25/40specificity25/30actionability15/3065 / 100
Rotate ossec.log daily into logs/ossec/YYYY/Mon/ossec-DD.log, then sign and compress it alongside alert logs.
— Describes new path and signing/compression but no reader action.4.5.0
0500
ossec-authd built against OpenSSL 4Not checked yetimproved60how completely this was documenteddepth25/40specificity25/30actionability10/3060 / 100
Build ossec-authd with OpenSSL 4, enforcing TLS 1.2 via TLS_method() and using ASNASNAutonomous System Number — a globally unique identifier assigned by regional internet registries to a network operator, letting routing protocols distinguish which organization controls a block of IP addresses..1 string accessors for certificate IP checks.
— Internal build/security change with no user-facing action.4.5.0
thinner coverage below
0600
Silent Windows installer skip flagsNot checked yetnew55how completely this was documenteddepth15/40specificity20/30actionability20/3055 / 100
Skip IIS logging and integrity checking in silent Windows agent installs with /IISLogging=no or /IntegrityChecking=no.
— Named installer flags but no example invocation shown.4.5.0
0700
notify_time scope extended to agent status checksNot checked yetimproved55how completely this was documenteddepth20/40specificity20/30actionability15/3055 / 100
Honor monitord.notify_time for agent_control status and remoted send eligibility, not only monitord disconnect alerts.
— Names the config key and affected components but gives no example.4.5.0
0800
IIS W3C log field extractionNot checked yetnew40how completely this was documenteddepth15/40specificity15/30actionability10/3040 / 100
Extract User-Agent, action, and srcport from default IIS W3C access logs.
— Bare list of extracted fields with no further detail.4.5.0
Trend Vision One shipped FIPSFIPSA set of US federal standards published by NIST that define approved cryptographic algorithms and modules; cyber tools use it to meet compliance requirements in government and regulated environments. mode control via new CLISHCLISHA shell-like CLI framework used in network and security appliances to expose a restricted, menu-driven command interface to operators, hiding the underlying OS while enforcing role-based command access. commands on the Service Gateway appliance, and full VPCVPCA logically isolated virtual network inside a public cloud provider's infrastructure, letting tools deploy resources with private addressing, controlled routing, and firewall rules without managing physical hardware. enforcement for File Security's AWS Lambda functions with validated, per-region SSM-stored configuration.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Trend Vision One is a unified cybersecurity platform that provides extended detection and response (XDR) across endpoints, networks, email, and cloud environments.
Trend Vision One shipped FIPSFIPSA set of US federal standards published by NIST that define approved cryptographic algorithms and modules; cyber tools use it to meet compliance requirements in government and regulated environments. mode control via new CLISHCLISHA shell-like CLI framework used in network and security appliances to expose a restricted, menu-driven command interface to operators, hiding the underlying OS while enforcing role-based command access. commands on the Service Gateway appliance, and full VPCVPCA logically isolated virtual network inside a public cloud provider's infrastructure, letting tools deploy resources with private addressing, controlled routing, and firewall rules without managing physical hardware. enforcement for File Security's AWS Lambda functions with validated, per-region SSM-stored configuration.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ WHAT SHIPPED ·
2 FEATURES0 OF 2 FEATURESmost completely described firstwhat's the number?
Each feature carries 0–100 for how completely the vendor documented
it — not how big or important the work is. A major capability
described in eight words scores low, and that is the finding.
depth0–40what it does and how it works or what changed
specificity0–30names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point
Code and config rank first by construction: a documented endpoint,
flag or config key scores at the top of specificity and actionability,
so it sorts above a feature described only in prose. Under
60 the notes go thin — everything below the “thinner
coverage” line is thinner documentation, not smaller work.
Hover any meter for that feature's three sub-scores.
0100
VPC enforcement for FSS Lambda functionsNot checked yetnew95how completely this was documenteddepth35/40specificity30/30actionability30/3095 / 100
File Security Storage stack version 4.5.4 for AWS lets you configure VPC placement for all FSS Lambda functions by providing VpcSubnetIDs, VpcSecurityGroupIDs, RegionalVpcSubnetIDs, and RegionalVpcSecurityGroupIDs together, or leaving all four empty to run outside a VPC.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄ Settings are now validated at deploy and update time to catch misconfigured stacks before Lambda functions launch, and per-region VPC configuration is stored as SSM parameters under /FSS/<StackSetName>/RegionalVpcConfig/ in every deployed region, removed automatically when the stack is deleted.
Retrieve the saved per-region VPC configuration from SSM after deployment to audit which subnets and security groups each region is using.
FIPS mode CLISH commands on Service GatewayNot checked yetnew45how completely this was documenteddepth25/40specificity10/30actionability10/3045 / 100
New CLISH commands let administrators show, enable, or disable FIPS mode on the Service Gateway appliance directly from the command line; changes prompt for confirmation and trigger an automatic reboot.
Cribl is the AI Platform for Telemetry, providing open, flexible infrastructure that helps enterprises manage, connect, and analyze their data across IT, security, and observability environments.
Privacy 3.1 raises recall by up to 13.85 and precision by up to 5.59 percentage points over 3.0 with no change in throughput or memory usage.
Updated detectionPrivacy 3.1 models improve recall by up to 13.85 percentage points and precision by up to 5.59 points over Privacy 3.0, with no change in throughput or memory usage. Not checked yet
New supported platformAn updated benchmark for model evaluation uses outlier-seeded test cases — including sensitive values in Windows event XML, heavily escaped JSON, headerless CSV, and free-form ticket text — with roughly three quarters of events drawn from log types never seen in training. Not checked yet
OXO is a security scanning orchestrator for the modern age.
oxo added a command for retrieving the logging service account when deploying it as an on-prem scanner.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ WHAT SHIPPED ·
1 FEATURE0 OF 1 FEATUREmost completely described firstwhat's the number?
Each feature carries 0–100 for how completely the vendor documented
it — not how big or important the work is. A major capability
described in eight words scores low, and that is the finding.
depth0–40what it does and how it works or what changed
specificity0–30names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point
Code and config rank first by construction: a documented endpoint,
flag or config key scores at the top of specificity and actionability,
so it sorts above a feature described only in prose. Under
60 the notes go thin — everything below the “thinner
coverage” line is thinner documentation, not smaller work.
Hover any meter for that feature's three sub-scores.
0100
Logging service account command for on-prem scannersNot checked yetnew40how completely this was documenteddepth15/40specificity10/30actionability15/3040 / 100
A new command retrieves the logging service account when running oxo as an on-prem scanner.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
— Names the capability but not the exact command syntaxv2.25.2
Varonis added flexible scheduling for automated searches, supporting quarterly and annual recurrence patterns.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Varonis connects data, AI, and threat detection into one complete security platform that allows companies to keep innovating, without increasing risk.
Varonis added flexible scheduling for automated searches, supporting quarterly and annual recurrence patterns.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ WHAT SHIPPED ·
1 FEATURE0 OF 1 FEATUREmost completely described firstwhat's the number?
Each feature carries 0–100 for how completely the vendor documented
it — not how big or important the work is. A major capability
described in eight words scores low, and that is the finding.
depth0–40what it does and how it works or what changed
specificity0–30names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point
Code and config rank first by construction: a documented endpoint,
flag or config key scores at the top of specificity and actionability,
so it sorts above a feature described only in prose. Under
60 the notes go thin — everything below the “thinner
coverage” line is thinner documentation, not smaller work.
Hover any meter for that feature's three sub-scores.
0100
Quarterly and annual search schedulingNot checked yetnew60how completely this was documenteddepth25/40specificity20/30actionability15/3060 / 100
Scheduled searches can now run at configurable month-based intervals — including quarterly and annual — with day-of-month or weekday-in-month execution options, eliminating manual re-runs of recurring reports.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
— Describes mechanism and options but no UI path or config keysnapshot-20261120
MIND introduces real-time outbound email DLPDLPData Loss Prevention: a category of controls that inspect and block sensitive data from leaving an organisation via email, web, or removable media, giving cyber tools a policy enforcement layer over data in motion and at rest. blocking for Microsoft Exchange Online, adding email as a new enforcement point alongside existing endpoint and browser policies.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
MIND data security platform puts data loss prevention (DLP) and insider risk management (IRM) on autopilot - automatically identify, detect & prevent data leaks
MIND introduces real-time outbound email DLPDLPData Loss Prevention: a category of controls that inspect and block sensitive data from leaving an organisation via email, web, or removable media, giving cyber tools a policy enforcement layer over data in motion and at rest. blocking for Microsoft Exchange Online, adding email as a new enforcement point alongside existing endpoint and browser policies.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ WHAT SHIPPED ·
1 FEATURE1 OF 1 FEATUREmost completely described firstwhat's the number?
Each feature carries 0–100 for how completely the vendor documented
it — not how big or important the work is. A major capability
described in eight words scores low, and that is the finding.
depth0–40what it does and how it works or what changed
specificity0–30names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point
Code and config rank first by construction: a documented endpoint,
flag or config key scores at the top of specificity and actionability,
so it sorts above a feature described only in prose. Under
60 the notes go thin — everything below the “thinner
coverage” line is thinner documentation, not smaller work.
Hover any meter for that feature's three sub-scores.
0101
Real-time email DLP blocking for Exchange OnlineNot checked yetnew69how completely this was documenteddepth32/40specificity22/30actionability15/3069 / 100
MIND now blocks outbound email in real time for Microsoft Exchange Online by routing mail through MIND for a policy decision before Exchange delivers it, requiring no agent installed on Exchange and no manual configuration in the Exchange admin center.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄ Email policy components are now available in MIND's policy builder alongside existing endpoint and browser policies, making email a third enforcement point on the platform. Because enforcement happens at the mail-flow level, sends from phones, tablets, and personal devices receive the same DLP protection as sends from managed workstations.
ConductorOne is an identity governance platform that automates workforce access management and deploys AI agents to streamline access reviews and provisioning.
ConductorOne's request forms now support conditional field visibility and regex-based answer validation, letting teams build smarter, more guided access request intake.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ WHAT SHIPPED ·
1 FEATURE0 OF 1 FEATUREmost completely described firstwhat's the number?
Each feature carries 0–100 for how completely the vendor documented
it — not how big or important the work is. A major capability
described in eight words scores low, and that is the finding.
depth0–40what it does and how it works or what changed
specificity0–30names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point
Code and config rank first by construction: a documented endpoint,
flag or config key scores at the top of specificity and actionability,
so it sorts above a feature described only in prose. Under
60 the notes go thin — everything below the “thinner
coverage” line is thinner documentation, not smaller work.
Hover any meter for that feature's three sub-scores.
0100
Regex validation on request form fieldsNot checked yetnew76how completely this was documenteddepth30/40specificity24/30actionability22/3076 / 100
Text and text area fields on request forms can be validated against a regular expression — such as a ticket ID or cost center code pattern — with a custom error message.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄ The web UI checks the answer as the requester types and Slack checks on submit, blocking submission until the answer matches.
Enforce a ticket ID format on a request form text field so free-text answers that don't match your ITSMITSMIT Service Management: a framework of practices for delivering and supporting IT services, covering ticketing, change control, and incident response. Cyber tools integrate with it to automate alert-to-ticket workflows. pattern are blocked before submission.
📍In the web UI, open a Text field on a request form, enable validation, enter a regular expression such as ^[A-Z]+-[0-9]+$, and add a custom error message like 'Enter a ticket ID in the format PROJ-1234'.
Secure users, AI agents, and more with Auth0, a production-ready authentication and authorization platform built to help you ship faster and drive sign-ups.
Auth0 Security Agent Early access adds guided triage and remediation: it recommends Attack Protection and Tenant ACLACLAn access control list, a set of rules allowing or denying operations for particular users, services or network addresses. It decides who can use a resource. changes and lets admins apply them after authenticating.
The Security Agent triages detected attacks and recommends Attack Protection and Tenant ACL configuration changes, letting an Auth0 tenant administrator review and apply each one after authenticating. Not checked yet
Re-run triage after applying recommendations to check the current status of an ongoing attack. Not checked yet
Okta Identity Engine is a secure, scalable, and highly available authentication and user management for any app.
Okta Identity Engine's 2026.47.1 release adds new audit logging for agent gateways, expands authentication options with on-demand identity verification and SPIFFESPIFFEAn open standard, maintained by the CNCF, for assigning cryptographic identities to workloads in dynamic infrastructure so services can authenticate each other without passwords or static credentials.JWT-SVIDJWT-SVIDA SPIFFE Verifiable Identity Document encoded as a JSON Web Token, used to prove workload identity across trust boundaries where X.509 certificates are impractical, such as HTTP headers. client auth, and previews new device-related visibility and mapping features.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ WHAT SHIPPED ·
5 FEATURES1 OF 5 FEATURESmost completely described firstwhat's the number?
Each feature carries 0–100 for how completely the vendor documented
it — not how big or important the work is. A major capability
described in eight words scores low, and that is the finding.
depth0–40what it does and how it works or what changed
specificity0–30names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point
Code and config rank first by construction: a documented endpoint,
flag or config key scores at the top of specificity and actionability,
so it sorts above a feature described only in prose. Under
60 the notes go thin — everything below the “thinner
coverage” line is thinner documentation, not smaller work.
Hover any meter for that feature's three sub-scores.
0101
SPIFFE JWT-SVID client authentication (GA)Not checked yetnew35how completely this was documenteddepth10/40specificity15/30actionability10/3035 / 100
Okta Identity Engine now supports SPIFFE JWTJWTAn open standard (RFC 7519) for encoding claims as a signed, optionally encrypted JSON object passed between parties. Cyber tools use it to verify identity and permissions without querying a central session store.-SVID client authentication, now generally available.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Device Visibility for WindowsNot checked yetnew35how completely this was documenteddepth15/40specificity15/30actionability5/3035 / 100
Okta Identity Engine adds Device Visibility for Windows, covering Desktop MFAMFAMulti-factor authentication, requiring proof from different categories, such as a password and a hardware key, to sign in. It reduces reliance on a password alone. and Device-Bound SSOSSOSingle sign-on, using one identity provider to sign in to multiple applications. It reduces separate passwords and centralizes account and access management., in Early Access.
Smart Card certificate policy AMR mappingNot checked yetnew25how completely this was documenteddepth10/40specificity10/30actionability5/3025 / 100
Okta Identity Engine adds Smart Card certificate policy AMRAMRAutonomous Mobile Robot — a self-navigating wheeled platform that maps its environment and plans its own path; cyber tools might target AMR firmware or fleet-management APIs in OT/ICS assessments. mapping, available in Early Access Preview.
Risk Ledger is a cutting-edge platform for real-time supply chain security, which enables organisations to Defend-as-One because every link matters in cybersecurity.
Risk Ledger adds an AI-assisted Domain Review Agent for all clients, PDF Review Reports on completed reviews, inline community signals in assessments, and a flexible re-review date picker.
Domain Review Agent is now on by default for all clients, running an AI first pass across all domain controls in a supplier assessment and surfacing flagged controls under an AI Findings on controls filter with a Findings (AI) tab per control. Not checked yet
Review Reports now auto-generate a PDF when a supplier review is completed, downloadable from the review drawer, containing the recommendation, approval decisions with who/what/when, non-compliant controls, open risks, remediations, full assessment answers, and optionally the AI supplier summary. Not checked yet
Community signals shared by peers now appear inline inside a supplier's assessment view, with a per-control icon and a filter to show only controls that carry a signal, for communities with review sharing switched on. Not checked yet
Re-review dates for completed reviews can now be adjusted from the supplier profile, with options from 1 month to 2 years or a custom date. Not checked yet
Decision makers now receive an email notification linking directly to the review when they are asked to decide; this can be turned off in notification settings. Not checked yet
+2 moreshow less
Review Status replaces Approval Status across the dashboard, supplier list, Insights, and saved views, showing completed reviews as Reviewed, Reviewed (Approved), or Reviewed (Rejected). Not checked yet
A regenerate button lets reviewers re-run the Domain Review Agent after a supplier updates their answers or compliance settings change, while previously dismissed findings stay hidden. Not checked yet
LogicGate launched GRCGRCGovernance, Risk, and Compliance — a framework for aligning IT controls with business policy, regulatory requirements, and risk management so organizations can audit and report on their posture systematically. Agents in Risk Cloud, bringing agentic AI into Third-Party Risk ManagementThird-Party Risk ManagementAssessing and monitoring the security risk an organization takes on through its vendors, suppliers and service providers. It covers the questionnaires, evidence and continuous checks behind letting a third party hold your data., AI GovernanceAI GovernancePolicies, standards, and controls that organizations use to manage risk, accountability, and compliance across AI systems — giving cyber tools a framework to audit model behavior and enforce acceptable-use boundaries., Enterprise Risk Management, and Business Continuity Management applications with full audit logging and configurable triggers.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
LogicGate Risk Cloud is the leading AI-powered GRC platform with 30+ applications for enterprise risk, third-party risk, cyber risk and compliance management.
LogicGate launched GRCGRCGovernance, Risk, and Compliance — a framework for aligning IT controls with business policy, regulatory requirements, and risk management so organizations can audit and report on their posture systematically. Agents in Risk Cloud, bringing agentic AI into Third-Party Risk ManagementThird-Party Risk ManagementAssessing and monitoring the security risk an organization takes on through its vendors, suppliers and service providers. It covers the questionnaires, evidence and continuous checks behind letting a third party hold your data., AI GovernanceAI GovernancePolicies, standards, and controls that organizations use to manage risk, accountability, and compliance across AI systems — giving cyber tools a framework to audit model behavior and enforce acceptable-use boundaries., Enterprise Risk Management, and Business Continuity Management applications with full audit logging and configurable triggers.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ WHAT SHIPPED ·
1 FEATURE0 OF 1 FEATUREmost completely described firstwhat's the number?
Each feature carries 0–100 for how completely the vendor documented
it — not how big or important the work is. A major capability
described in eight words scores low, and that is the finding.
depth0–40what it does and how it works or what changed
specificity0–30names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point
Code and config rank first by construction: a documented endpoint,
flag or config key scores at the top of specificity and actionability,
so it sorts above a feature described only in prose. Under
60 the notes go thin — everything below the “thinner
coverage” line is thinner documentation, not smaller work.
Hover any meter for that feature's three sub-scores.
0100
GRC Agents for TPRM, AI Governance, ERM, BCMNot checked yetnew80how completely this was documenteddepth35/40specificity27/30actionability18/3080 / 100
Risk Cloud now includes GRC Agents that can act within Third-Party Risk Management, AI Governance, Enterprise Risk Management, and Business Continuity Management applications.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄ Every agent action is captured in the same Record History and Field History audit trail used for human users, including confidence score, reasoning, and source citations. Agent runs can be triggered manually or automatically based on a workflow signal, configured per application, and each agent operates under its own agent user identity governed by the same permission sets as human users, which admins can enable or disable at any time. Agent inputs and outputs are processed through OpenAI via API, retained for up to 30 days for safety monitoring, and are not used to train models.
Review what a GRC Agent changed on a vendor intake record — open the audit trail immediately after an agent run to verify every field, link, and record the agent touched.
📍In Risk Cloud, open the vendor intake record the agent processed, then navigate to Record History and Field History to see each change the agent made, its confidence score, reasoning, and source citations.
Compyl is the AI-powered GRC platform that maps one control library to SOC 2, ISO 27001, HIPAA, PCI DSS and 70+ frameworks, so you stay audit-ready year-round.
Compyl adds custom fields on Controls (admin-defined, editable in drawer and grid, importable via CSV) in this release.
Admins can now define custom fields on Controls; users can view and edit them in the control drawer, see them as columns in the Controls grid, and populate them via CSV import. Not checked yet
Runlayer is an MCP gateway that enforces identity-mapped permissions and detects threats across the agents and MCP servers employees use.
Runlayer adds per-virtual-key LLMLLMA large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. Gateway routing, dynamic subagents, an agent kill switch, audit-logged grant refusals, tools/list policy enforcement, sort_by=recommended on the agents API, and HTML artifact publishing.
Give a virtual key its own LLM Gateway routing rules and fallbacks using runlayer_llm_gateway_usage_rule with scope = "virtual_key" and runlayer_llm_gateway_routing_rule in Terraform, or set both in the key's Edit dialog under Settings → LLM Gateway → Virtual keys. Not checked yet
Manage MDMMDMMobile device management, centrally configuring and managing enrolled devices. Administrators use it to apply settings, distribute applications and enforce device security policies. configurations as code with the new runlayer_mdm_configuration Terraform resource in the next Terraform provider release. Not checked yet
Agents on the Pi harness can now start dynamic subagents using the new Agent tool, with read-only, planning, or general-purpose workers that run in the background and use the agent's connectors, skills, and identity, with at most two running at once. Not checked yet
Sort agents by recent organizational usage with sort_by=recommended on the agents API and the list_agentsMCPMCPModel Context Protocol, a standard for connecting AI applications to external tools and data sources. A compatible server exposes capabilities an assistant can use through a common interface. tool; the response also includes users_30d, runs_30d, and last_used_at per agent. Not checked yet
An agent harnessagent harnessThe software around an AI agent that manages its tools, execution loop, state and limits. It determines how the model's proposed actions are run and checked. can now request connector tools in a flat list by sending the X-Runlayer-Tool-Exposure: force-flat header, including on agent account endpoints and Plugins with Dynamic tools; force-nested keeps clients on search_tools and execute_tool. Not checked yet
+13 moreshow less
Session search is now available to every organization with Sessions on the Sessions page and via the Runlayer MCP session search tools, with no separate opt-in required. Not checked yet
Organizations with session topics now see a Session topics section on the Analytics page. Not checked yet
New integrationTurn off a model for all users in Settings → AI Providers by toggling the Enabled switch on any model under Model access, hiding it from all agent model pickers and rejecting new picks. Not checked yet
The agent kill switch now stops agent tokens, OBO tokens, API keys, LLM gateway keys, upstream connector calls, and sandbox file operations when an agent is disabled; turning it back on restores all grants. Not checked yet
The Runlayer Assistant now hands off work to connected-app agents (such as Notion) when used in web chat, Go, iMessage, or Microsoft Teams, as it already did in Slack. Not checked yet
Agents can now publish a folder with an index.html as an artifact, with scripts, styles, images, fonts, and media combined into one page; screenshot_artifact previews the same page. Not checked yet
A refused attempt to grant an agent or agent account more tools or resources than the person's own policy allows is now recorded in the audit log as agent_grant_refused, naming who tried and what was refused. Not checked yet
Connector-condition policies (meta.server.*) are now applied in tools/list, omitting tools denied by connector-condition rules from the list rather than listing them and refusing at call time. Not checked yet
The Google Workspace connector can now convert .pptx to Google Slides, .docx to a Google Doc, and .xlsx.xlsxThe file format used by Microsoft Excel for spreadsheets, based on the Open XML standard. Cyber tools use it to ingest or export structured tabular data in a format analysts already work with. to a Google Sheet when creating files. Not checked yet
The identity rebind API can now reset a user's sign-in binding and release the directory user held by a deactivated duplicate, letting a Super Admin resolve these cases without Runlayer support. Not checked yet
Super Admins can now see a user's directory user ID and refused sync matches on the user's Manage tab, view all refused matches under Settings → Sign-in & provisioning, and look up the directory user for an email through the API. Not checked yet
The sidebar is now an icon rail with a section panel; Home, Inbox, connectors, skills, assets, Teams, and Settings open from the rail, with new manage pages for connectors, skills, and endpoints for admins. Not checked yet
Reviewers can now re-run a draft procedure's benchmark directly from the review drawer when it has failed or is waiting for a replay, without waiting for a new version. Not checked yet
!Stdio connectors that run inside the Runlayer backend now run on Python 3.14 and start in /tmp instead of /app; a uv script whose requires-python excludes 3.14 no longer starts. If a connector relied on /app, set cwd in its config.
!A uv stdio connector that runs its own project now builds a fresh environment on each start; set UV_ISOLATED=0 in its config to reuse the project's .venv again.
!Any --python <version> or -p <version> argument must be removed from a uv stdio connector's args — ClickHouse connectors added from the catalog before this release include --python 3.13 and won't start until it's removed.
!Sending plain text for Google Slides via the Google Workspace connector now fails with an error message instructing the caller to send a .pptx instead; previously it silently created a Google Doc. Existing Google Drive connectors need to be redeployed to receive file conversion.
!The seen_by_user_count API field on connector tools now stops at 50 for tools seen by 50 or more users.
!The deprecated connections API no longer widens an existing connection.
!The session filter on the Audit Logs page is now offered only to organizations with Sessions.
Ifixai is an independent Auditing of AI Agents. Run by human or the agent itself, to answer the most crucial question in the AI Agent Economy.
ifixai's v4.4.0 release adds a full self-check CLI with SARIFSARIFStatic Analysis Results Interchange Format, a standard for sharing code-analysis findings, including locations and rule details. It lets scanners report results to a common review interface. and PR-comment output, trace import from five observability platforms, CI integrations for GitHub ActionsGitHub ActionsGitHub's built-in CI/CD system that runs automated workflows triggered by repository events, letting cyber tools automate testing, building, and releasing without managing separate pipeline infrastructure./GitLab/pre-commitpre-commitAn open-source framework for managing and running git hooks, maintained by the pre-commit project, that lets cyber tools enforce checks automatically on every commit before code reaches a repository., and a Pro audit offer that now surfaces by default and can auto-install the iFixAi MCPMCPModel Context Protocol, a standard for connecting AI applications to external tools and data sources. A compatible server exposes capabilities an assistant can use through a common interface. server across 14 coding agents.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ WHAT SHIPPED ·
6 FEATURES1 OF 6 FEATURESmost completely described firstwhat's the number?
Each feature carries 0–100 for how completely the vendor documented
it — not how big or important the work is. A major capability
described in eight words scores low, and that is the finding.
depth0–40what it does and how it works or what changed
specificity0–30names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point
Code and config rank first by construction: a documented endpoint,
flag or config key scores at the top of specificity and actionability,
so it sorts above a feature described only in prose. Under
60 the notes go thin — everything below the “thinner
coverage” line is thinner documentation, not smaller work.
Hover any meter for that feature's three sub-scores.
0100
Agent self-check via `ifixai audit`Not checked yetnew93how completely this was documenteddepth35/40specificity28/30actionability30/3093 / 100
ifixai audit <agent-url> runs the smoke suite and any pinned findings against an OpenAI-shaped chat endpoint to self-check an agent; pass --yes to confirm the sandbox prompt, set IFIXAI_AGENT_KEY for the agent's key, and use --judge-provider mock to skip external calls.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄ Exit codes are now defined: 0 means nothing failed, 1 means a check or pinned finding failed, and 2 means no verdict was reached.
— Full command, flags, env var and exit codes givenv4.4.0
0201
Pro audit offer and MCP server auto-install across coding agentsNot checked yetbreaking93how completely this was documenteddepth35/40specificity30/30actionability28/3093 / 100
ifixai offer prints the free Pro audit offer for a coding agent; pass --no-promo or set IFIXAI_NO_PROMO=1 to skip it. ifixai offer --add-server adds the iFixAi MCP server (https://mcp.ifixai.ai/mcp) to every coding agent found on the machine, covering Claude Code, Codex, Gemini CLI, GitHub Copilot CLI, Devin, Cline, Amp, opencode, Cursor, VS Code Copilot, Kiro, Junie, Warp and Antigravity. As of this release every real ifixai run with a person at the terminal now opens with this offer by default; picking the Pro audit ends the run before any call to the agent and exits 0.
— Exact command, URL, env var and full list of 14 agents namedv4.4.0
0300
Trace import from five observability platformsNot checked yetnew83how completely this was documenteddepth30/40specificity28/30actionability25/3083 / 100
ifixai traces import ingests trace data from OpenTelemetryOpenTelemetryAn open standard and set of tools for collecting traces, metrics and logs from applications. It lets services export observability data without tying instrumentation to one monitoring vendor. GenAI, Langfuse, LangSmith, Arize/OpenInference and Braintrust exports into a trajectory capture; input over 64 MiB or 100,000 events is refused rather than truncated.
— Command, five named sources and explicit size limitsv4.4.0
0400
CI integrations for self-check in GitHub Actions, GitLab CI and pre-commitNot checked yetnew83how completely this was documenteddepth30/40specificity28/30actionability25/3083 / 100
actions/self-check runs the self-check in GitHub Actions, uploads SARIF to code scanning, and can post or update a verdict comment. ci/gitlab/ifixai-self-check.gitlab-ci.yml runs it in GitLab CI, keeping the report and SARIF as artifacts and optionally posting the verdict on the merge request. The .pre-commit-hooks.yaml pre-commit integration runs the self-check on push and checks the regressions file on commit.
— Three named integration files/paths with exact behavior describedv4.4.0
0500
Findings explain, validate, SARIF and PR comment commandsNot checked yetnew81how completely this was documenteddepth25/40specificity28/30actionability28/3081 / 100
ifixai findings explain explains a self-check failure in plain language (what the agent was asked, what it should have done, and the framework controls it covers); ifixai findings validate-regressions checks that .ifixai/regressions.yaml is well-formed; ifixai findings sarif writes SARIF 2.1.0SARIF 2.1.0A JSON-based file format, standardized by OASIS, for expressing static analysis results so that cyber tools from different vendors can share findings through a single, machine-readable schema. output from self-check results; and ifixai findings comment writes a pull-request comment.
— Four commands named with exact syntax and output formatv4.4.0
thinner coverage below
0600
Default model updates for Gemini and Bedrock providersNot checked yetbreaking45how completely this was documenteddepth15/40specificity20/30actionability10/3045 / 100
The Gemini default model changed from gemini-2.0-flash to gemini-3.6-flash, and the Bedrock default model changed to global.anthropic.claude-sonnet-4-6.
— Names exact model identifiers but no migration guidancev4.4.0
└──▷ BREAKING ON UPGRADE
!
Every real ifixai run with a person at the terminal now opens with the free Pro audit offer; pass --no-promo or set IFIXAI_NO_PROMO=1 to skip it, and picking the Pro audit ends the run before any call to your agent and exits 0.
!
The Gemini default model changed from gemini-2.0-flash to gemini-3.6-flash.
!
The Bedrock default model changed to global.anthropic.claude-sonnet-4-6.
Joplin v3.7.24 introduces a first-class OpenAI provider built on the Responses API for AI-assisted notes, alongside a set of end-to-end encryption safeguards including key fingerprint validation and a breaking change disabling unencrypted uploads by default.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Joplin - the privacy-focused note taking app with sync capabilities for Windows, macOS, Linux, Android and iOS.
Joplin v3.7.24 introduces a first-class OpenAI provider built on the Responses API for AI-assisted notes, alongside a set of end-to-end encryption safeguards including key fingerprint validation and a breaking change disabling unencrypted uploads by default.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ WHAT SHIPPED ·
2 FEATURES1 OF 2 FEATURESmost completely described firstwhat's the number?
Each feature carries 0–100 for how completely the vendor documented
it — not how big or important the work is. A major capability
described in eight words scores low, and that is the finding.
depth0–40what it does and how it works or what changed
specificity0–30names real surfaces — APIs, flags, formats, limits, numbers
actionability0–30enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point
Code and config rank first by construction: a documented endpoint,
flag or config key scores at the top of specificity and actionability,
so it sorts above a feature described only in prose. Under
60 the notes go thin — everything below the “thinner
coverage” line is thinner documentation, not smaller work.
Hover any meter for that feature's three sub-scores.
0101
E2EE security safeguards and key validationNot checked yetbreaking55how completely this was documenteddepth25/40specificity18/30actionability12/3055 / 100
Joplin adds a UI to validate public key fingerprints when sharing notebooks, letting users confirm the identity of share recipients; now asks for confirmation before disabling end-to-end encryption to prevent accidental deactivation; detects invalid public-private key pairs and allows recovering from them; and disables unencrypted uploads by default when E2EEE2EEEnd-to-end encryption: a scheme where data is encrypted on the sender's device and decrypted only by the intended recipient, so no intermediary server can read it. is enabled, a breaking change for setups that relied on unencrypted uploads alongside E2EE.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
— Describes what changed but no exact UI path or flagv3.7.24
0200
OpenAI provider via Responses APINot checked yetnew40how completely this was documenteddepth15/40specificity15/30actionability10/3040 / 100
Joplin adds a first-class OpenAI provider built on the Responses API for AI-assisted note features.
— Names the Responses API but no setup steps or config keyv3.7.24
└──▷ BREAKING ON UPGRADE
!
Unencrypted uploads are now disabled by default when E2EE is enabled; existing setups that relied on unencrypted uploads alongside E2EE will need to review their sync configuration.
Was this useful?
No tools match. Clear the filter and try again.
my-toolchain — 0 tools
Know when your tools change
Add what you run. We’ll show what each tool shipped since your version, and which breaking changes to check first.
Copy this, send it to your AI assistant, then come back with its reply.
Nothing is sent from here. You paste the reply in the next step.
1 Copy
2 Paste
3 Confirm
Paste the reply
One tool per line is best. Versions are optional, and any format works. Names we don’t track are sent to us, without versions, so we can consider adding them.