The daily firehose — everything the toolchain shipped today, already filtered.
We read every release from the 1315 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.
Hacktron adds finding IDs, activity timelines, Jira/Linear status sync, Whitebox retest wizards, and Azure DevOps Remediation across this snapshot.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ HOW TO FIND IT
Keep bot PRs like Dependabot in scope for review while excluding a specific bot author by name.
📍In your repository configuration, add the bot's handle to exclude.authors to prevent it from consuming a seat while all other bot pull requests remain eligible for review.
Let Hacktron open and track fix pull requests on an Azure DevOps repository as part of an automated Remediation workflow.
📍In the Hacktron web app, go to the finding detail for an Azure DevOps-hosted repository and start Remediation — Hacktron will open and track the fix pull request on Azure DevOps as it does on GitHub and GitLab.
Filter the findings list to specific findings by pasting two or more finding IDs into the search box, with Source, Reopened, and Date range filters each saved in the URL.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Finding status now syncs with Jira and Linear — admins enable or disable each sync direction separately in the integration settings, Jira updates follow your project's workflow transitions, and triaging a finding in the web app, Slack, a pull request comment, or the API moves the linked Linear issue to the matching workflow state.
Every member can now create, rotate, and revoke their own API keys directly from the unified Personal settings page, without admin involvement.
Bot pull requests (such as Dependabot and Renovate) are now reviewed on GitHub; use exclude.authors to exclude specific bots.
Request a retest of a completed Whitebox pentest from the scans list or scan detail page; the retest covers every reported finding except false positives and accepted risk.
+5 moreshow less
Every finding now has a short, stable, human-readable ID that appears in the URL, the finding row, panel, and full page and can be copied by clicking it, and can carry an OWASP Top 10:2025OWASP Top 10:2025A ranked list of the ten most critical web application risk categories, maintained by the Open Worldwide Application Security Project, used by cyber tools to prioritize detection, testing, and remediation coverage. category that is settable from the overflow menu and filterable in the findings list.
The finding panel now has an Activity tab showing full history: engine detection, consecutive sightings collapsed into Detected again in N scans, and every status or severity edit with actor and changed fields.
The command palette now leads with a contextual action group for the open finding: change status, change severity, mark or unmark as duplicate, create a ticket, start Remediation, copy its ID, copy its link, and open the full page.
Whitebox PDF reports now include repositories and targets in scope, testing window, severity rubric, limitations, confidentiality notice, and methodology and coverage detail — the Methodology section shows coverage by area (functions selected per area) and the Coverage Detail table caps the unselected entry points list at 50 rows.
Findings with recorded live-validation evidence now include a Live Validation section in the PDF report showing what was established, what was not, and any blocker.
Filter the findings list to specific findings by pasting two or more finding IDs into the search box, with Source, Reopened, and Date range filters each saved in the URL.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Finding status now syncs with Jira and Linear — admins enable or disable each sync direction separately in the integration settings, Jira updates follow your project's workflow transitions, and triaging a finding in the web app, Slack, a pull request comment, or the API moves the linked Linear issue to the matching workflow state.
Every member can now create, rotate, and revoke their own API keys directly from the unified Personal settings page, without admin involvement.
Request a retest of a completed Whitebox pentest from the scans list or scan detail page; the retest covers every reported finding except false positives and accepted risk.
Every finding now has a short, stable, human-readable ID that appears in the URL, the finding row, panel, and full page and can be copied by clicking it, and can carry an OWASP Top 10:2025OWASP Top 10:2025A ranked list of the ten most critical web application risk categories, maintained by the Open Worldwide Application Security Project, used by cyber tools to prioritize detection, testing, and remediation coverage. category that is settable from the overflow menu and filterable in the findings list.
+2 moreshow less
The finding panel now has an Activity tab showing full history: engine detection, consecutive sightings collapsed into Detected again in N scans, and every status or severity edit with actor and changed fields.
Findings with recorded live-validation evidence now include a Live Validation section in the PDF report showing what was established, what was not, and any blocker.
Xalgorix autonomous AI pentesting agents — real-time reconnaissance, vulnerability detection, and exploitation orchestration. Go + TypeScript.
xalgorix now defaults XALGORIX_CONVERGE_FINISH and XALGORIX_CONVERGE_RECONCILE to true, automatically closing scans once coverage obligations are met︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ GET THIS VERSION
$ git clone --branch v4.6.183 https://github.com/xalgorix/xalgorix.git
# already have the repo? check out this version:$ git checkout v4.6.183
└──▷ TRY IT
Restore the old indefinite-scan behaviour in a CI environment where you want the scan to run until manually stopped.
Scans now close automatically after XALGORIX_CONVERGE_FINISH_IDLE iterations (default 500) of no new validated findings once the coverage threshold is cleared, with high-confidence open leads getting a XALGORIX_CONVERGE_FINISH_LEAD_WINDOW-iteration (default 100) verification window before closure; set XALGORIX_CONVERGE_FINISH=false to restore the previous behaviour.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Content-discovery artifacts such as VCSVCSVersion control system: software that tracks changes to files over time, letting teams collaborate and roll back mistakes. Cyber tools use it to manage source code, configs, and signatures. metadata, config and secret files, logs, and OS cruft now owe only information-exposure checks rather than every injection class.
Scans now close automatically after XALGORIX_CONVERGE_FINISH_IDLE iterations (default 500) of no new validated findings once the coverage threshold is cleared, with high-confidence open leads getting a XALGORIX_CONVERGE_FINISH_LEAD_WINDOW-iteration (default 100) verification window before closure; set XALGORIX_CONVERGE_FINISH=false to restore the previous behaviour.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
!XALGORIX_CONVERGE_FINISH now defaults to true; scans that previously ran indefinitely past their coverage threshold will now terminate automatically after idle iterations.
!XALGORIX_CONVERGE_RECONCILE now defaults to true; exploratory plan tasks without a vulnerability class that previously kept a fully covered scan open indefinitely will now close once all measurable obligations are done.
3 more releases in this issue
· 2026-11-05 → 2026-11-07
xalgorix adds NVIDIA NeMo Switchyard as a tool-trajectory-aware router strategy, configurable via switchyard_router_config YAML, with per-session state, auditable spend logs, and custom tool-class vocabularies.
Configure Switchyard routing via switchyard_router_config in model_list YAML, specifying capable and efficient model names, with an optional 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. or decision-model judge for inconclusive turns.
Connect Claude Code to a Switchyard router by setting ANTHROPIC_BASE_URL, ANTHROPIC_AUTH_TOKEN, and ANTHROPIC_MODEL to the router name, with no client-side code changes required.
xalgorix now supports NVIDIA NeMo Switchyard as an auto_router/switchyard strategy that routes each agent turn based on the session's rolling tool-call history — escalating to a capable model on failures and stepping back down on recovery — rather than on prompt content alone.
Custom tool vocabularies let internal or 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. tools be declared as observe, mutate, plan, or new to contribute to Switchyard's trajectory signals alongside built-in classifications.
Switchyard routing decisions are recorded in spend logs with the selected model and evidence, enabling cost and quality analysis per session.
Use auto-updating -latest model aliases — such as openai/gpt-luna-latest, anthropic/claude-sonnet-latest, or deepseek/deepseek-pro-latest — in the Python SDK and AI Gateway so the model behind the name updates without touching application code.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
xalgorix adds model deprecation alerts, scheduled email reminders, and per-key usage tracking to help prevent outages during model migrations︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Enable model deprecation email reminders by turning on Model Deprecations under Logging & Alerts → Email Alerts → Email Notifications in the UI; xalgorix schedules two reminders at 30 days and 1 day before the provider's deprecation date.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Review per-key traffic during a migration using the Review impact view, which shows a daily requests chart broken down by key alongside affected keys, teams, owners, and tags.
View model deprecation alerts — including provider, model, deprecation date, requests in the last 30 days, and last-seen traffic — under Observability → Alerts → Model deprecations in the UI.
Mark a migration complete with Mark resolved so the alert moves to Resolved; new requests to the old model automatically reopen it.
ZeroPath is an AI-native SAST platform that finds 2x more real vulnerabilities with 75% fewer false positives.
ZeroPath adds directory team sync, secret exposure snapshots, Gerrit CI scanning, restricted service keys, and new CLI commands for contexts, rules, and tags︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ TRY IT
List all scan contexts as JSON to pipe into a downstream script or audit report.
$ zeropath contexts list --json
Start a Gerrit change scan from CI on self-hosted deployments by calling scans.scanGerritChange; switch each server between Discovery: Polling and Discovery: CI-triggered in Settings.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Manage scan context, natural-language rules, and repository tags with the new CLI commands zeropath contexts, zeropath rules, and zeropath tags, each supporting list, get, create, update, and delete subcommands plus a --json flag for scripting.
Retrieve a secret finding's exposure snapshot — including who can read the repository on GitHub or GitLab and which ZeroPath users can open the finding — via the new issues.exposureSnapshot API.
Members with API Token Create can now create restricted service keys scoped to selected repositories or all repositories, edit a key's permissions with Edit permissions without rotating it, and only organization admins can grant Full access.
Map identity-provider directory groups to ZeroPath teams under Settings by enabling Enable directory team mappings once a directory is connected.
+5 moreshow less
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. exports now carry valid SPDX license expressions, with licenses absent from the SPDX License List marked NOASSERTION.
Dependency scans and dependency monitoring now create Jira tickets for findings that meet the auto-export score threshold under Full scans and dependency monitoring, and resolving or dismissing a finding in ZeroPath updates its ticket.
Notification settings can now be scoped to specific repositories under Specific Repositories when created.
The dashboard's Total Open card now shows its count broken down by status, including Needs Review, and each status filters the Issues list to that status.
Issue search now returns each finding's lane and evidence, supports filtering by lane and dependency signals, and supports sorting by priority; issue list and search results also carry linked tickets.
Start a Gerrit change scan from CI on self-hosted deployments by calling scans.scanGerritChange; switch each server between Discovery: Polling and Discovery: CI-triggered in Settings.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Manage scan context, natural-language rules, and repository tags with the new CLI commands zeropath contexts, zeropath rules, and zeropath tags, each supporting list, get, create, update, and delete subcommands plus a --json flag for scripting.
Retrieve a secret finding's exposure snapshot — including who can read the repository on GitHub or GitLab and which ZeroPath users can open the finding — via the new issues.exposureSnapshot API.
Members with API Token Create can now create restricted service keys scoped to selected repositories or all repositories, edit a key's permissions with Edit permissions without rotating it, and only organization admins can grant Full access.
Map identity-provider directory groups to ZeroPath teams under Settings by enabling Enable directory team mappings once a directory is connected.
+1 moreshow less
Dependency scans and dependency monitoring now create Jira tickets for findings that meet the auto-export score threshold under Full scans and dependency monitoring, and resolving or dismissing a finding in ZeroPath updates its ticket.
Mend Platform is a traditional application security misses AI risk. Standalone AI security tools miss your code.
Mend Platform adds an AI assistant for natural-language risk queries and granular scan-category controls in Scan Profiles︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ HOW TO FIND IT
Reduce false positives from noisy scan categories by excluding File Read and Database Read from your Scan Profile before the next scheduled scan.
📍In the Mend Platform, go to Settings › Scan Profile, expand the low-probability scan categories tree, uncheck File Read and Database Read, then save — the count indicator confirms how many categories remain active. Changes apply on the next scan.
Mend Platform now lets users selectively include or exclude specific low-probability scan categories — such as File Read or Database Read — within the Scan Profile, replacing the previous binary toggle with a configurable selection tree that shows a count of selected categories.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Mend Platform adds the Mend Intelligent Agent, an AI assistant that accepts natural-language questions to explore platform capabilities, navigate to relevant views, and generate executive-level risk summaries across all applications and projects; enabled by default for organizations with a signed AI consent addendum.
Clover Security automates security design reviews and threat modelling from specs, flags design-to-implementation drift, and feeds security context to coding agents.
Clover Security adds a Notion Custom Agent template that runs a security review inline on a spec page in about a minute︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ HOW TO FIND IT
Run a security review on a spec before any coding agent turns it into code, without leaving Notion.
📍1. Add the Clover Spec Security Review template to your Notion workspace and connect it to your Clover account.
2. Open your spec page in Notion.
3. Type @Clover Spec Security Review on the page.
4. Clover reads the design and writes a Clover Security Notes block inline on the page.
Adds Clover Spec Security Review, a Notion Custom Agent template that lets you trigger a security review by typing @Clover Spec Security Review on any spec page, writing a Clover Security Notes block inline.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Security notes written into the Notion spec are read by coding agents alongside the design, so security requirements shape generated code from the first line.
Get started with Semgrep to help you catch, flag, and fix real vulnerabilities before they ship.
Semgrep adds self-serve Guardian trials, Codex and PowerShell support, Detect-2Shell workflow, Bun lockfile parsing, and AWS CodeArtifact private registry resolution︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ HOW TO FIND IT
Trigger the Detect-2Shell workflow to find all Node.js and PHP runtime versions across your repositories and identify vulnerable or unpinned ones.
📍In the Semgrep platform, go to Agentic Workflows, select the Detect-2Shell workflow, choose your target projects, and click Run.
Allow Supply Chain to resolve private Maven or Gradle packages from AWS CodeArtifact when scanning Java projects with transitive dependency tracking.
📍In the Semgrep platform, go to Settings › Supply Chain › Private Registries, add an AWS CodeArtifact registry entry, and configure the IAM role ARN for authentication.
Semgrep Guardian now supports Codex with automatic code, dependency, and secret scanningsecret scanningChecking files, code or other data for exposed credentials such as passwords, API keys and tokens. Findings help teams remove the secret and revoke or rotate it. — no Semgrep CLI install needed.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Guardian hooks now run on PowerShell in addition to Bash, covering developers on Windows.
Guardian now alerts when a compromised version of the Claude Code extension for Cursor containing an SSH backdoor is detected.
Guardian updates itself automatically on new releases, with each download verified to SLSA Level 3SLSA Level 3A supply-chain integrity framework tier defined by the OpenSSF requiring that builds run on a hardened, ephemeral build platform and that provenance is generated and signed by the build service itself, not the developer..
In Semgrep Agentic Workflows, the new Detect-2Shell workflow finds Node.js and PHP versions across Docker images, .nvmrc files, infrastructure as code, and CI — marking each as vulnerable, unpinned, fixed, or unknown and using AI to resolve versions set in variables or base images — and Agentic Workflows can now be scheduled to run daily, weekly, or monthly, requiring Unified Policies, the default for new organizations.
+5 moreshow less
Supply Chain now parses bun.lock, giving Bun projects dependency, malware, and vulnerability coverage.
Every package in requirements.txt, pinned or not, now appears in the dependency inventory and 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..
New organizations can start a 30-day free trial of Guardian and upgrade in the product without a sales call.
AI provider settings now offer Automatic and Manual modes, Cloudflare Workers AI, and a direct connection to Anthropic with your own API key.
The Wiz integration now sends findings from Azure DevOps, Bitbucket Cloud, and Bitbucket Data Center repositories in addition to GitHub and GitLab.
Semgrep Guardian now supports Codex with automatic code, dependency, and secret scanningsecret scanningChecking files, code or other data for exposed credentials such as passwords, API keys and tokens. Findings help teams remove the secret and revoke or rotate it. — no Semgrep CLI install needed.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Guardian hooks now run on PowerShell in addition to Bash, covering developers on Windows.
Guardian now alerts when a compromised version of the Claude Code extension for Cursor containing an SSH backdoor is detected.
Guardian updates itself automatically on new releases, with each download verified to SLSA Level 3SLSA Level 3A supply-chain integrity framework tier defined by the OpenSSF requiring that builds run on a hardened, ephemeral build platform and that provenance is generated and signed by the build service itself, not the developer..
In Semgrep Agentic Workflows, the new Detect-2Shell workflow finds Node.js and PHP versions across Docker images, .nvmrc files, infrastructure as code, and CI — marking each as vulnerable, unpinned, fixed, or unknown and using AI to resolve versions set in variables or base images — and Agentic Workflows can now be scheduled to run daily, weekly, or monthly, requiring Unified Policies, the default for new organizations.
+2 moreshow less
Supply Chain now parses bun.lock, giving Bun projects dependency, malware, and vulnerability coverage.
New organizations can start a 30-day free trial of Guardian and upgrade in the product without a sales call.
OWASP dep-scan is a next-generation security and risk audit tool based on known vulnerabilities, advisories, and license limitations for project dependencies. Both local repositories and container images are supported as the input, and the tool is ideal for integration.
dep-scan adds --fail-on-error to abort scans on failures and DEPSCAN_SERVER_TRUSTED_HOST_NAMES to allowlist server hostnames︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Abort a scan immediately on failures with the new --fail-on-error flag.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Allowlist server hostnames beyond localhost by setting the new DEPSCAN_SERVER_TRUSTED_HOST_NAMES environment variable; without an API key, the Host header must match localhost, an IP address, or a name in this list.
2 details in Everything
└──▷ BREAKING ON UPGRADE
!--server-allowed-paths no longer blocks scans that generate a BOMBOMA Bill of Materials listing every component, library, and dependency in a software project; cyber tools use it to track what is in a codebase and flag known-vulnerable components quickly..
!On Windows, a custom batch file set through CDXGEN_CMD or a similar variable no longer receives arguments containing &, |, <, >, ^, ", % or !.
!Without DEPSCAN_SERVER_API_KEY, the Host header must be localhost, an IP address, or a name listed in DEPSCAN_SERVER_TRUSTED_HOST_NAMES; a server reached by a non-loopback name or address accepts path and url only in a JSON body.
!A path parameter must be an existing file or directory and cannot look like an option, URL or package URL; on Windows, UNC and device paths are refused.
Arnica is an agentic application security platform that secures AI-generated code with SAST, SCA, IaC, secrets detection, agent governance, and automated remediation.
Arnica adds Agentic Pentesting to test live applications and confirm whether findings are exploitable using code, cloud, runtime, and business context︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Arnica now includes Agentic Pentesting, a new capability that tests live applications to confirm or debunk whether SASTSASTStatic application security testing, analyzing source code or compiled code for vulnerabilities without running the application. It helps developers find problems before deployment., AI SAST, 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., secrets, and IaCIaCInfrastructure as Code: the practice of defining and provisioning compute infrastructure through machine-readable config files rather than manual processes, giving cyber tools repeatable, version-controlled environment setup. findings are actually exploitable at runtime.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
A private connector lets agents test internal apps behind a firewall by making an outbound connection to Arnica — no VPNVPNA virtual private network, an encrypted tunnel between devices or networks. It carries traffic across another network as though the endpoints had a private connection. required, with only allow-listed apps reachable.
Agents work from a context graph built from four sources — source code (SAST, AI SAST, SCA, secrets, IaC), cloud (AWS, Azure, GCP, Kubernetes, containers), runtime (live traffic, DASTDASTDynamic application security testing, probing a running application from the outside for vulnerabilities. It checks behavior exposed through the application's interfaces.), and business context (architecture diagrams, design docs, threat models) — to surface risks invisible to any single scanner.
Playbooks let teams customize the agent's scope and priorities per application, using built-in playbooks or writing their own.
A live browser view lets practitioners watch agent activity in real time and take over at any point during a test run.
+2 moreshow less
Each finding receives a verdict — confirmed, lowered, debunked, flagged for human review, or newly discovered — with the steps the agent took recorded for triage.
Runs execute in an isolated sandbox on Arnica's infrastructure with a per-run budget the team sets.
MobSF is an automated mobile application pen-testing, malware analysis and security assessment framework for Android, iOS and Windows apps, with static and dynamic analysis.
CVE-2026-68922: Arbitrary File Read via Path Traversal in ZIP Uploads (CVSSCVSSCommon Vulnerability Scoring System, a method for describing a vulnerability's characteristics and severity. Its score is not the probability that attackers will exploit it. 5.5, affects <= v4.4.6)︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
16 more releases in this issue
· 2024-03-22 → 2026-07-05
CVE-2026-68923: Regression: CSRFCSRFCross-site request forgery, an attack that makes a signed-in browser send an unwanted request to a site. The site may accept the request using the victim's existing session. checks not enforced after Django migration (CVSSCVSSCommon Vulnerability Scoring System, a method for describing a vulnerability's characteristics and severity. Its score is not the probability that attackers will exploit it. 6.5, affects <= 4.5.0)︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
CVE-2026-68924: Zip Bomb Denial of Service via Per-File Size Limit Bypass in ZIP/APK Extraction (CVSSCVSSCommon Vulnerability Scoring System, a method for describing a vulnerability's characteristics and severity. Its score is not the probability that attackers will exploit it. 4.9, affects <= 4.5.0)︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
CVE-2026-33545: SQL Injection in SQLite Database Viewer utils (CVSSCVSSCommon Vulnerability Scoring System, a method for describing a vulnerability's characteristics and severity. Its score is not the probability that attackers will exploit it. 5.3, affects <= 4.4.5)︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
CVE-2026-24490: Stored XSSXSSCross-site scripting, an attack that runs malicious script in another user's browser through a trusted site. It can steal data or perform actions using the user's session. via Manifest Analysis - Dialer Code Host Field (CVSSCVSSCommon Vulnerability Scoring System, a method for describing a vulnerability's characteristics and severity. Its score is not the probability that attackers will exploit it. 8.1, affects 4.4.4)︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
CVE-2025-58161: Path Traversal in GET /download/<filename> using absolute filenames in MobSF data directory (CVSSCVSSCommon Vulnerability Scoring System, a method for describing a vulnerability's characteristics and severity. Its score is not the probability that attackers will exploit it. 4.3, affects 4.4.0)︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
CVE-2025-58162: Arbitrary File Write (AR-Slip) via Absolute Path in .a Extraction (CVSSCVSSCommon Vulnerability Scoring System, a method for describing a vulnerability's characteristics and severity. Its score is not the probability that attackers will exploit it. 6.5, affects 4.4.0)︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
CVE-2025-46730: Web Server Resource Exhaustion via ZIP of Death Attack (CVSSCVSSCommon Vulnerability Scoring System, a method for describing a vulnerability's characteristics and severity. Its score is not the probability that attackers will exploit it. 6.8, affects <=4.3.2)︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
CVE-2025-46335: Stored Cross Site Scripting (XSSXSSCross-site scripting, an attack that runs malicious script in another user's browser through a trusted site. It can steal data or perform actions using the user's session.) via malicious SVG Icon Upload (CVSSCVSSCommon Vulnerability Scoring System, a method for describing a vulnerability's characteristics and severity. Its score is not the probability that attackers will exploit it. 5.4, affects <=4.3.2)︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
CVE-2025-31116: SSRFSSRFServer-side request forgery, an attack that tricks a server into making requests chosen by an attacker. Those requests can reach internal services the attacker cannot access directly. on assetlinks_check with DNS Rebinding (CVSSCVSSCommon Vulnerability Scoring System, a method for describing a vulnerability's characteristics and severity. Its score is not the probability that attackers will exploit it. 4.4, affects 4.3.1)︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
CVE-2025-24804: Partial Denial of Service due to strict regex check in iOS report view URL (CVSSCVSSCommon Vulnerability Scoring System, a method for describing a vulnerability's characteristics and severity. Its score is not the probability that attackers will exploit it. 4.3, affects 4.3.0)︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
CVE-2025-24805: Local Privilege escalationPrivilege escalationGaining permissions beyond those an account or process initially had. It can turn limited access into administrator control or access to another user's resources. due to leaked REST APIREST APIA web API design style where clients interact with resources over HTTP using standard methods like GET, POST, and DELETE, giving cyber tools a predictable, language-agnostic way to integrate with external services. key in web UI (CVSSCVSSCommon Vulnerability Scoring System, a method for describing a vulnerability's characteristics and severity. Its score is not the probability that attackers will exploit it. 5.5, affects 4.3.0)︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
CVE-2024-53999: Stored Cross-Site Scripting Vulnerability in Recent Scans Diff or Compare (CVSSCVSSCommon Vulnerability Scoring System, a method for describing a vulnerability's characteristics and severity. Its score is not the probability that attackers will exploit it. 8.1, affects <=4.2.8)︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
CVE-2024-43399: Zip Slip Vulnerability in .a Static Library Files (CVSSCVSSCommon Vulnerability Scoring System, a method for describing a vulnerability's characteristics and severity. Its score is not the probability that attackers will exploit it. 8.0, affects <=4.0.6)︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
CVE-2024-41955: Open Redirect in Login Redirect (CVSSCVSSCommon Vulnerability Scoring System, a method for describing a vulnerability's characteristics and severity. Its score is not the probability that attackers will exploit it. 5.2, affects <=4.0.4)︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
CVE-2024-31215: SSRFSSRFServer-side request forgery, an attack that tricks a server into making requests chosen by an attacker. Those requests can reach internal services the attacker cannot access directly. in firebase database check (CVSSCVSSCommon Vulnerability Scoring System, a method for describing a vulnerability's characteristics and severity. Its score is not the probability that attackers will exploit it. 6.3, affects <=3.9.7)︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
CVE-2024-54000: Potential Fix Bypass - [SSRFSSRFServer-side request forgery, an attack that tricks a server into making requests chosen by an attacker. Those requests can reach internal services the attacker cannot access directly. Vulnerability on assetlinks_check(act_name, well_knowns)] (CVSSCVSSCommon Vulnerability Scoring System, a method for describing a vulnerability's characteristics and severity. Its score is not the probability that attackers will exploit it. 7.5, affects <=3.9.6)︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Enable static shellcode analysis on CAPE-extracted payloads or directly submitted shellcode by setting enabled, cape_payloads, and submitted_shellcode under [static_shellcode_check] in the processing config.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
CAPEv2 now runs shellcode analysis on extracted and submitted shellcode, surfacing framework family detection via the new shellcode_utils integration module.
Hrtng is an IDA Pro plugin with a rich set of features: decryption, deobfuscation, patching, lib code recognition and various pseudocode transformations
hrtng adds the Malware Analysis Add-on for IDA Pro 9.5, bringing automated unpacking, full-system emulation, and sandbox trace overlay into the disassembly workflow︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ USE IT
Unpack a malware sample headlessly in an idalib script so that downstream string extraction or code indexing operates on the inner payload rather than the packer stub.
python
start_emulator_with_backend("speakeasy")
# IDA will emulate, detect OEP, rebuild IAT, and update the IDB automatically
# Then use ida_dbg APIs for further stepping or breakpoint control
Invoke the full-system Windows emulator headlessly from an idalib script or AI agent with start_emulator_with_backend('speakeasy'), then control it via the existing ida_dbg APIs.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Adds automated malware unpacking that integrates Bochs, Sogen, and Speakeasy emulator backends alongside remote physical hardware and virtual machine snapshot backends, rebuilding the IAT and recovering the original entry point (OEP) in-place.
Adds sandbox trace overlay that pulls VMRay analysis archives directly from VMRay's API and displays API calls, parameters, return values, and code coverage inline in the disassembly and pseudocode views.
Adds a 'Sandbox > Select events in this function' context menu action to highlight all sandbox trace rows whose addresses fall within the currently viewed function.
Adds an Apply all API names/types in function context menu action to resolve indirect calls by applying names, parameter types, and return types from the sandbox trace into both the disassembly and pseudocode views.
+2 moreshow less
Adds the ability to import unmapped memory regions (such as shellcode or injected payloads) directly into the IDB by double-clicking an unmapped address row in the sandbox trace view.
Adds integrations with VirusTotal, MalwareBazaar, and VMRay to surface existing reporting on a sample during triage.
Fickling is a Python pickling decompiler and static analyzer
fickling adds detection of private stdlib imports, subinterpreter gadgets, and underscore attributes, and now hooks pickle._load, pickle._loads, and pickle._Unpickler︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ GET THIS VERSIONpipxgit
pipx
$ pipx install fickling==0.1.13
# already installed? upgrade to this version:$ pipx install --force fickling==0.1.13
git
$ git clone --branch v0.1.13 https://github.com/trailofbits/fickling.git
# already have the repo? check out this version:$ git checkout v0.1.13
PyPI attestation · checked 8 Nov
Detect private stdlib imports, including subinterpreter modules that can execute arbitrary code, in scanned pickle files.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Extend private import detection to underscore attributes and qualified names in pickle files.
Hook pickle._load, pickle._loads, and pickle._Unpickler so that fickling's safety checks also cover Python-internal pickle entrypoints.
Analyze PyTorch checkpoints with PyTorch's own ZIP reader to reduce risks from parsing differentials between fickling and PyTorch.
Reject inconsistent FRAME boundaries in pickle streams that could hide executable opcodes from analysis, and recompute frame lengths when serializing edited pickles.
+2 moreshow less
Expand the unsafe imports blocklist with additional code execution, file access, and networking gadgets.
Add Python 3.15 support; the minimum supported version is now Python 3.11.
Detect private stdlib imports, including subinterpreter modules that can execute arbitrary code, in scanned pickle files.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Hook pickle._load, pickle._loads, and pickle._Unpickler so that fickling's safety checks also cover Python-internal pickle entrypoints.
Reject inconsistent FRAME boundaries in pickle streams that could hide executable opcodes from analysis, and recompute frame lengths when serializing edited pickles.
4 more in Everything
└──▷ BREAKING ON UPGRADE
!Python 3.10 is no longer supported; the minimum is now Python 3.11.
!The max_acceptable_severity parameter has been removed from FicklingContextManager; any code that passes it will break.
Ecapture is a capture SSL/TLS plaintext with eBPF—no MITM proxy or custom CA installation. Supports Linux and Android on x86_64\arm64 and loong64.
ecapture adds QUICQUICA UDP-based transport protocol standardized by the IETF that bakes in TLS 1.3, multiplexed streams, and connection migration, giving cyber tools low-latency encrypted transport without head-of-line blocking. v1/v2 verification, GnuTLSGnuTLSAn open-source TLS/SSL library maintained by the GNU Project that implements secure communications protocols, giving cyber tools a portable, LGPL-licensed alternative to OpenSSL for encrypted connections. 3.8.10–3.8.13 support, and live-decryption options for HTTP/1.1, HTTP/2HTTP/2The second major version of the HTTP protocol, standardized by the IETF, adding multiplexed streams, header compression, and binary framing over a single TCP connection to cut latency and reduce overhead., and HTTP/3HTTP/3The third major version of HTTP, standardized by the IETF, which runs over QUIC instead of TCP, eliminating head-of-line blocking and reducing connection latency for faster, more resilient data transfer.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ GET THIS VERSION
$ git clone --branch v2.10.0 https://github.com/gojue/ecapture.git
# already have the repo? check out this version:$ git checkout v2.10.0
└──▷ TRY IT
Capture and reconstruct full HTTP/2 and HTTP/1.1 messages from live TLS traffic on port 443, writing structured events to a file instead of stdout.
$ sudo ecapture tls -m text --text-source=pcapng-i eth0 --eventaddr /tmp/http.events 'tcp port 443'
Inspect decrypted MQTT 3.1.1 traffic inside TLS by selecting the MQTT application parser alongside live pcapng decryption.
$ sudo ecapture tls -m text --text-source=pcapng--text-protocol mqtt -i eth0 'tcp port 8883'
Capture a Go TLS client that upgrades SMTPSMTPThe application-layer protocol used to send email between servers and from clients to servers; cyber tools use it to relay alerts, phishing simulations, or exfiltrated data over standard mail infrastructure. to TLS via STARTTLSSTARTTLSA protocol command that upgrades a plaintext connection to an encrypted TLS session without changing ports, used in SMTP, IMAP, and LDAP so tools can negotiate encryption opportunistically on standard service ports., saving the raw pcapng for later analysis while streaming events to stdout.
$ sudo ecapture gotls -m text --text-source=pcapng--text-tls-upgrade smtp --elfpath /usr/bin/go-mailer -i eth0 --pcapfile /tmp/raw.pcapng 'tcp port 587'
Reconstruct complete HTTP/1.1 and HTTP/2 messages from live captured traffic using -m text --text-source=pcapng with the tls, gotls, or gnutls subcommands; output goes to --eventaddr (stdout by default).︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Filter reconstructed output to raw TLS/QUIC stream bytes with --text-raw, or to incremental HTTP/1.x, HTTP/2, HTTP/3, SSESSEServer-Sent Events, a browser API and HTTP-based protocol where a server pushes a stream of text events to a client over a single long-lived connection, giving tools a lightweight alternative to WebSockets for one-way real-time updates., and HTTP/2 gRPCgRPCAn open-source remote procedure call framework from Google that uses HTTP/2 and Protocol Buffers to let services call each other's functions across a network, giving tools low-latency, strongly typed inter-process communication. events with --text-http-events.
Limit per-message gzip expansion with --text-max-decoded-body N (requires --text-http-events; zero keeps content encoding).
Declare a STARTTLS/SSLRequest/CONNECT negotiation with --text-tls-upgrade smtp|imap|pop3|ldap|postgres|mysql|connect|https-connect; https-connect includes one inner TLS layer.
Parse DTLSDTLSA variant of TLS adapted for UDP, standardized by the IETF, that adds handshake reliability and replay protection without sacrificing datagram semantics — letting cyber tools encrypt low-latency or connectionless traffic. 1.2 ECDHE-RSARSAThe oldest widely used public-key algorithm, for both encryption and signatures, with security resting on factoring large numbers. A large quantum computer would break it. AES-128/256-GCM records over UDPUDPA network protocol that sends individual messages without guaranteeing delivery or order. Applications trade built-in reliability for less overhead and can add their own recovery. with --text-udp-protocol dtls (default UDP protocol is QUIC).
+2 moreshow less
ecapture now supports GnuTLS versions 3.8.10 through 3.8.13 struct layouts.
ecapture now verifies QUIC v1 and v2 capture in the gotls module.
Reconstruct complete HTTP/1.1 and HTTP/2 messages from live captured traffic using -m text --text-source=pcapng with the tls, gotls, or gnutls subcommands; output goes to --eventaddr (stdout by default).︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Filter reconstructed output to raw TLS/QUIC stream bytes with --text-raw, or to incremental HTTP/1.x, HTTP/2, HTTP/3, SSESSEServer-Sent Events, a browser API and HTTP-based protocol where a server pushes a stream of text events to a client over a single long-lived connection, giving tools a lightweight alternative to WebSockets for one-way real-time updates., and HTTP/2 gRPCgRPCAn open-source remote procedure call framework from Google that uses HTTP/2 and Protocol Buffers to let services call each other's functions across a network, giving tools low-latency, strongly typed inter-process communication. events with --text-http-events.
Limit per-message gzip expansion with --text-max-decoded-body N (requires --text-http-events; zero keeps content encoding).
Declare a STARTTLS/SSLRequest/CONNECT negotiation with --text-tls-upgrade smtp|imap|pop3|ldap|postgres|mysql|connect|https-connect; https-connect includes one inner TLS layer.
Parse DTLSDTLSA variant of TLS adapted for UDP, standardized by the IETF, that adds handshake reliability and replay protection without sacrificing datagram semantics — letting cyber tools encrypt low-latency or connectionless traffic. 1.2 ECDHE-RSARSAThe oldest widely used public-key algorithm, for both encryption and signatures, with security resting on factoring large numbers. A large quantum computer would break it. AES-128/256-GCM records over UDPUDPA network protocol that sends individual messages without guaranteeing delivery or order. Applications trade built-in reliability for less overhead and can add their own recovery. with --text-udp-protocol dtls (default UDP protocol is QUIC).
Sysdig Secure is a cloud-native security platform that detects threats, monitors runtime activity, and scans containers and Kubernetes environments for vulnerabilities and compliance issues.
Sysdig Secure adds AI Defense, a runtime security and governance layer for AI agents, 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. servers, and local models across workstations, Linux, and Kubernetes︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Sysdig AI Defense now discovers every AI agent, MCP server, skill, and plugin running across employee workstations, production Linux, Kubernetes, and managed agent platforms — including unapproved and shadow instances — and ties each to the account and person behind it.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Block, Confirm, and Kill process policy responses now intercept an agent's tool call before it runs or end a process the moment it starts, with the agent receiving the policy name and an exception link.
Adds a new Explorer view that expands each identity into the departments, devices, harnesses, MCP servers, models, skills, and threats connected to it, and a new Threats view that shows behavior chains, the sessions behind a detected incident, and the identity the agent acted under.
Session replay now steps through every tool call in a flagged session in order, with each detection naming the rule, policy, identity, harness, and model that fired.
Sysdig AI Defense now rates each person's model tier against the demand of their work by team, flagging light work on expensive models and heavy work on cheap ones.
+3 moreshow less
Managed Falco rules for AI agent risks — covering MCP servers and tools, secrets and data, and destructive commands — are now available as readable, editable Falco rules.
Every AI Defense event can now be forwarded to a SIEMSIEMSecurity information and event management, collecting logs from many systems and correlating events to find suspicious activity. Analysts use it to investigate alerts and incidents. alongside existing security evidence.
All AI Defense records are written to the customer's Sysdig tenant off the device, in the customer's chosen region, with prompts and file edits stripped before leaving the machine.
Give a build step its own private cache directory so concurrent builds do not share or corrupt each other's cache.
$ buildah build --layers-f - . <<'EOF'
FROM golang:1.23
RUN --mount=type=cache,target=/go/pkg/mod,sharing=private go mod download
EOF
Add a --umask flag to buildah run to control the umask for the executed command.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Adds --include to buildah add and buildah copy as the complement to the existing --exclude flag, and --required-path (which must be used with --include) to assert that specific paths were actually copied.
Allow the --checksum option in ADD Containerfile instructions and in buildah add, buildah build, and buildah copy to accept Git commit SHAs (or a prefix) for Git sources.
Cache mounts now accept sharing=private, giving a command its own exclusive cache directory.
Suppress Containerfile step output when --log-level error or --log-level fatal is passed to buildah build or buildah push.
+4 moreshow less
Docker-compatible --keep-git-dir support is now available for ADD of Git sources; without it the .git directory is excluded by default.
The buildah info command now includes information about system CDI (container device interface) settings.
The buildah pull command now recognises --policy=newer as equivalent to --policy=ifnewer.
Builds now honor containers.confhttp_proxy and env settings during image builds, aligning behavior with podman run.
Add a --umask flag to buildah run to control the umask for the executed command.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Adds --include to buildah add and buildah copy as the complement to the existing --exclude flag, and --required-path (which must be used with --include) to assert that specific paths were actually copied.
Docker-compatible --keep-git-dir support is now available for ADD of Git sources; without it the .git directory is excluded by default.
6 more in Everything
└──▷ BREAKING ON UPGRADE
!An ADD Containerfile command from Git sources no longer includes the .git directory by default; --keep-git-dir or --keep-git-dir=true is now required to include it.
Databricks is a cloud-based data lakehouse platform for data engineering, analytics, and machine learning built on Apache Spark.
Databricks now defaults new jobs compute to Auto access mode, switching to Standard unless ML runtime, GPUGPUA graphics processing unit, a processor that performs many calculations in parallel. AI workloads use GPUs to speed up model training and inference., or Runtime < 14.3 is selected︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ BREAKING ON UPGRADE
!New jobs compute created in the UI defaults to Auto access mode instead of Dedicated; workspaces with the compliance security profile enabled are affected in December 2026. Existing jobs are not changed, and a cluster policy enforcing a different access mode still takes precedence.
Microsoft Security is a find official documentation, practical know-how, and expert guidance for builders working and troubleshooting in Microsoft products.
Microsoft Security adds Dynamic Permissions for agent task-scoped access, Microsoft Execution Containers for local AI agent containment, and On-Premises Baseline Security Mod︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Dynamic Permissions now gives AI agents authority scoped to their current task, with Dynamic Agent Risk assessing whether the agent's authority aligns with changing behavior and Dynamic Conditional AccessConditional AccessA policy enforcement framework, most commonly Microsoft's in Entra ID, that grants or blocks resource access based on signals like user identity, device state, and location — letting cyber tools trigger step-up auth or deny risky sessions. narrowing permissions, restricting resource access, or requiring human approval for higher-impact actions.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Microsoft Execution Containers (MXC), now generally available, applies OS-enforced least-privilege containment for AI agents on Windows across files, networks, processes, and user interactions, with additional isolation options for higher-risk scenarios and upcoming integration with Microsoft Intune, Microsoft Entra, and Microsoft Agent 365.
Microsoft Intune Endpoint Privilege Management now supports system settings elevation, allowing standard users to change IPv4, IPv6IPv6The current version of the Internet Protocol, standardized by the IETF, using 128-bit addresses to replace the exhausted IPv4 space. Cyber tools target it because modern networks run dual-stack or IPv6-only., and DNS server values on network adapters without broad administrator rights, with administrator-controlled requirements for business justification or Windows authentication before changes are applied.
Secure Now now pairs information about recent threats with focused initiatives across security domains in Microsoft Security Exposure Management, connecting observed attack patterns to foundational security controls teams can prioritize.
Emergency Mitigation Service 2.0 for Exchange Server and Skype for Business Server, planned for December 2026, will enable the service by default and expand the security mitigations it can apply, with extension to SharePoint Server planned for January 2027.
+1 moreshow less
Hardened default configurations for SharePoint Server, Exchange Server, and Skype for Business Server, with configuration drift identification, are planned for Q1 2027.
Dynamic Permissions now gives AI agents authority scoped to their current task, with Dynamic Agent Risk assessing whether the agent's authority aligns with changing behavior and Dynamic Conditional AccessConditional AccessA policy enforcement framework, most commonly Microsoft's in Entra ID, that grants or blocks resource access based on signals like user identity, device state, and location — letting cyber tools trigger step-up auth or deny risky sessions. narrowing permissions, restricting resource access, or requiring human approval for higher-impact actions.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Microsoft Execution Containers (MXC), now generally available, applies OS-enforced least-privilege containment for AI agents on Windows across files, networks, processes, and user interactions, with additional isolation options for higher-risk scenarios and upcoming integration with Microsoft Intune, Microsoft Entra, and Microsoft Agent 365.
Microsoft Defender is Microsoft's security suite that protects devices, identities, and cloud resources against malware, phishing, and other threats.
Microsoft Defender adds a consolidated Defender Experts MDRMDRManaged detection and response, an outsourced service that monitors security signals, investigates threats and helps respond to incidents. It combines specialist analysts with detection tools. overview page and local AI agent/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. discovery on Windows and macOS; enhanced SQL agent now available in Azure Government.
Defender Experts MDR now offers a consolidated overview page with updated navigation and visibility into active and completed hunts, findings, and items requiring follow-up.
New detectionMicrosoft Defender now discovers more than 50 supported local AI agents and their configured model context protocols (MCPs) on Windows and macOS, surfacing agent, device, account, and MCP context together for security teams.
New supported platformThe enhanced agent for Defender for SQL Servers on machines is now generally available for resource- and subscription-level enablement in Azure Government cloud, using existing SQL infrastructure without requiring the Azure Monitor Agent.
NinjaOne is an improve IT efficiency with an autonomous patch management solution. Automate updates and secure endpoints across your organization with NinjaOne.
NinjaOne 16.0 adds a technician mobile app, real-time device chat for Windows, and early-access self-service OS updates for end users.
NinjaOne Assist now includes a technician-facing interface alongside the existing end-user experience in the same app, letting technicians handle remote support and device management from their phones.
NinjaOne Chat lets technicians initiate real-time chat with the person logged in to a managed Windows device, replacing ticket back-and-forth with in-context messaging.
New integrationEarly Access self-service OS updates let end users install policy-approved OS updates from the NinjaOne End User Portal without opening a ticket or waiting for a maintenance window.
NinjaOne Vulnerability Management 16.0 adds consolidated patch scanning, cutting scan time by ~90%, and expands third-party patch coverage to 146 new products︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Third-party patch catalog adds 146 new products (55 for Windows, 91 for macOS) added from April through September 2026, with an average of 12.3 patches published per day over the last 30 days.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
NinjaOne 16.0 adds unified browser policy configuration covering startup pages, bookmarks, HTTPSHTTPSHTTP carried over an encrypted TLS connection. It protects traffic in transit and lets a browser verify the server's certificate. enforcement, credential storage, notifications, and browser AI agent controls︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ HOW TO FIND IT
Roll out a compliance-focused browser baseline that blocks password saving, disables payment info storage, and enforces HTTPS across a pilot group before expanding org-wide.
📍In the NinjaOne console, go to Browser Management › Policies, create a new policy, enable HTTPS enforcement, set password storage to disabled, set payment info storage to disabled, and assign the policy to your pilot device group.
Standardize startup pages and bookmarks for new hires so their browser opens to internal tools with the right links already pinned — no per-machine scripting required.
📍In the NinjaOne console, go to Browser Management › Policies, configure startup pages and homepage URLs, add your bookmark list with pinned entries, and push the policy to your onboarding device group.
Set an organization-wide policy for built-in browser AI agents before they proliferate, to prevent unmanaged access to page content across managed devices.
📍In the NinjaOne console, go to Browser Management › Policies, locate the browser AI agent controls section, configure the allowed or restricted state, and apply the policy to all managed devices.
NinjaOne Browser Management now controls startup pages and homepages, letting admins define what opens when the browser launches and where the home button points, all within a single browser policy.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
NinjaOne Browser Management now controls bookmarks, allowing admins to manage bookmark toolbar behavior, pin bookmarks, and push custom bookmark lists to users across the organization.
NinjaOne Browser Management now enforces HTTPS connections across managed devices via browser policy, replacing reliance on user awareness.
NinjaOne Browser Management now sets policy on password saving, payment info storage, and cookie handling across all managed devices from a single policy.
NinjaOne Browser Management now controls browser notification permissions, letting admins decide whether sites can prompt users with notifications.
+1 moreshow less
NinjaOne Browser Management now includes browser AI agent controls, letting admins set policy for built-in browser AI features alongside other browser settings in the same policy.
NinjaOne Browser Management now controls startup pages and homepages, letting admins define what opens when the browser launches and where the home button points, all within a single browser policy.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
NinjaOne Browser Management now includes browser AI agent controls, letting admins set policy for built-in browser AI features alongside other browser settings in the same policy.
CyberArk is a protect your organization with Idira®. Modernize identity governance, automate access reviews, and secure every identity across your hybrid cloud environment.
CyberArk's Idira platform adds X.509X.509A standard defining the format of public-key certificates, maintained by the ITU-T, used by cyber tools to authenticate identities and establish encrypted connections over TLS and similar protocols.SVIDsSVIDsShort-lived, cryptographically verifiable identity documents issued by the SPIFFE framework to workloads, giving cyber tools a standard way to authenticate services to each other without managing certificates manually., automatic OAuth registration for 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. servers, Google Agent Identity attestation, AWS/Kubernetes node attestation, and Azure management group onboarding for Secrets Hub.
New supported platformSecure Workload Access now issues X.509-SVIDs alongside 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.-SVIDs, with trust bundles published in PEM, DER, or 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. bundle format and configurable certificate lifetime per trust domain — enabling mutual TLSmutual TLSMutual TLS, an encrypted connection where both client and server present certificates to prove their identity. It helps services verify who they are communicating with. and AWS IAMAWS IAMAWS Identity and Access Management is Amazon's service for controlling who and what can call AWS APIs, by attaching policies that allow or deny actions on resources. Cyber tools use it to scope credentials to least privilege. Roles Anywhere access without a separately managed CA.
Two new SWA node attestation methods — aws_iid (AWS Instance Identity Document on EC2EC2Amazon's cloud virtual-machine service, letting tools run on-demand servers without managing physical hardware.) and k8s_jwt (projected service account token on Kubernetes) — join the existing x509pop and k8s_psat methods, removing the need for pre-provisioned bootstrap credentials.
New integrationA new gemini_enterprise attestor on the SWA Server verifies a Google Agent Identity token and issues a 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. or X.509-SVIDX.509-SVIDAn X.509 certificate formatted to carry a SPIFFE Verifiable Identity Document, binding a cryptographic identity to a workload so cyber tools can authenticate services without passwords or static secrets. for Gemini Enterprise agents, requiring no SWA Agent on the host and no pre-placed key material.
New supported platformSecure AI Agents now automatically registers OAuth apps for MCP servers that support dynamic client registration (DCR), managing credential rotation so agents connect through the Identity Broker without holding MCP server credentials.
Unmanaged agents discovered in the AI agents inventory can now be remediated in a guided flow — registered as a new managed agent (pre-filled from discovery metadata) or linked to an existing managed registration — clearing the risk finding within minutes.
+3 moreshow less
New integrationSecrets Hub now onboards Azure management groups directly, creating a disabled secret store for each Key Vault in the group for bulk or API-driven enablement, without requiring individual secret store setup or full Microsoft Entra tenant onboarding.
New integrationSecrets Hub's AWS organization onboarding template, redeployed through Connect cloud environments, now automatically creates and removes secret stores as AWS regions change in local policies, switches authentication to AWS External ID, and restricts network access to specific static IPs.
The Idira Identity gateway now returns HTTP 429 for requests exceeding API rate limits.
!Legacy tenant role-based authentication for Secrets Hub's AWS organization onboarding is deprecated and will be removed in February 2027; redeploying the template through Connect cloud environments migrates to AWS External ID automatically.
OpenSSH 10.6p2 adds the post-quantumpost-quantumCryptography designed to stay secure against a large quantum computer, which could break today's RSA and elliptic-curve keys. NIST standardized the first algorithms in 2024 as ML-KEM, ML-DSA and SLH-DSA. hybrid signature scheme ssh-mldsa44-ed25519 as an allowed CA signature algorithm and adds PubkeyAuthOptions max-pk-ok:nnn to cap per-key︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
OpenSSH now supports ssh-mldsa44-ed25519 as an allowed CA signature algorithm, enabling post-quantum hybrid signatures for certificate authorities.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Zscaler is a cloud-based security platform that provides secure internet access, zero-trust network access, and threat protection for users and applications.
Zscaler's Zero Trust Browser now streams thick-client applications like SAP GUIGUIA graphical user interface, controls such as windows, buttons and menus. It lets people use software visually rather than entering commands. and developer IDEs through isolated, ephemeral cloud environments, removing the need for VDIVDIVirtual Desktop Infrastructure: a model where desktop operating systems run as VMs on a central server and stream to thin clients, giving cyber tools a single managed environment to monitor or scan. i︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Zscaler's Zero Trust Browser now supports thick-client application delivery through Application Isolation, running apps like SAP GUI, database utilities, and developer IDEs inside isolated, ephemeral cloud environments rendered directly in a browser tab.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Users access thick-client desktop environments from the Zero Trust Browser landing page alongside existing web and private applications, with no local VDI client, endpoint agent, or customer-managed infrastructure required.
Session persistence is enabled by default, allowing users to disconnect and reconnect to a running session with unsaved work intact.
Copy/paste and file downloads from isolated thick-client sessions are governed by policy, keeping enterprise data inside the cloud environment.
Zscaler's Zero Trust Browser now supports thick-client application delivery through Application Isolation, running apps like SAP GUI, database utilities, and developer IDEs inside isolated, ephemeral cloud environments rendered directly in a browser tab.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Session persistence is enabled by default, allowing users to disconnect and reconnect to a running session with unsaved work intact.
Copy/paste and file downloads from isolated thick-client sessions are governed by policy, keeping enterprise data inside the cloud environment.
Veeam Backup & Replication is backup, recovery, and replication software for virtual, physical, cloud, and SaaS workloads.
Veeam Backup & Replication v13.1.2 adds automatic best-practice repository settings on Data Domain add, plus Instant VM Recovery direct from deduplicated storage.
Updated supported platformNew compliance frameworkVeeam Backup & Replication now automatically applies recommended best-practice repository settings when you add a Dell Data Domain system as a backup repository, reducing manual configuration.
Veeam Backup & Replication can now drive DD Boost Managed File Replication between Data Domain systems, replicating backup files appliance-to-appliance in deduplicated form for offsite copies while tracking those copies for restore.
Updated integrationDell Data Domain (DD3410 and all higher current-line models) now holds the Veeam Ready Repository Primary Target certification, enabling Instant VM Recovery to run directly from Data Domain without a landing zone, as of Veeam Backup & Replication v13.1.2.
Warden is the secure gateway connecting AI agents to enterprise systems.
Warden adds an agent identity registry with sponsorship, governed-only roles, an OIDCOIDCOpenID Connect, a sign-in protocol built on OAuth 2.0. An application uses identity information from a trusted provider instead of maintaining a separate password for each user. auth method for operators, and a deny mode for unclaimed agents︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ TRY IT
Register a new agent, attach a team sponsor, and place it under governance so it can hold roles through static groups.
Create a static group granting specific roles to governed agents, then claim the agent's native 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. identity so virtual groups no longer apply to it.
After registering and sponsoring all known agents and verifying the audit log is quiet, switch the namespace to deny mode so any unclaimed agent is refused before role evaluation.
$ warden identity entity list -ungoverned
# Review audit log until quiet, then:
warden identity config write -unknown-aliases=deny
Operators can now sign in through their company identity provideridentity providerA service that authenticates users and supplies identity information to other applications. It is the sign-in authority behind many single sign-on systems. using the new oidc auth method with warden login, holding roles through virtual groups over IdPIdPA service that authenticates users and supplies identity information to other applications. It is the sign-in authority behind many single sign-on systems. claims.
Agent discovery now runs under namespace built-in policies and is audited.
Agents can call Microsoft GraphMicrosoft GraphMicrosoft's unified REST API for accessing data and services across Microsoft 365, Azure AD, and related cloud products, giving cyber tools a single authenticated endpoint to query users, emails, files, and sign-in logs. as the signed-in person they act for through Entra on-behalf-of, and Azure apps can authenticate with a certificate instead of a secret.
Amazon Bedrock now works keyless, with no model API key stored anywhere.
Token and credential caches are now sized in real bytes with a configurable budget.
Operators can now sign in through their company identity provideridentity providerA service that authenticates users and supplies identity information to other applications. It is the sign-in authority behind many single sign-on systems. using the new oidc auth method with warden login, holding roles through virtual groups over IdPIdPA service that authenticates users and supplies identity information to other applications. It is the sign-in authority behind many single sign-on systems. claims.
Agents can call Microsoft GraphMicrosoft GraphMicrosoft's unified REST API for accessing data and services across Microsoft 365, Azure AD, and related cloud products, giving cyber tools a single authenticated endpoint to query users, emails, files, and sign-in logs. as the signed-in person they act for through Entra on-behalf-of, and Azure apps can authenticate with a certificate instead of a secret.
Amazon Bedrock now works keyless, with no model API key stored anywhere.
2 more in Everything
└──▷ BREAKING ON UPGRADE
!Auth mounts no longer bind roles — any configuration relying on this will break on upgrade.
!Agent mounts now require an audience — existing mounts without one will fail on upgrade.
!There are twelve breaking changes in total; the upgrade guide must be read before upgrading from v0.21.0.
Vanta is a compliance automation platform that continuously monitors a company's cloud, code, and identity systems to help achieve and maintain security certifications like SOC 2, ISO 27001, and HIPAA.
Vanta adds 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.-based controls management, a native DocuSign integration, rebuilt Jira/Linear/Vercel connectors, and 27 new risk library scenarios︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Delete custom controls through the Vanta Agent or MCP, with associated mappings removed under the same authorization rules as the web app.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Through authorized MCP-connected agents, update common control fields, retrieve common control subdomain information, retrieve common control details and current status, retrieve deactivated control details without changing status (also via the Vanta Agent), and deactivate common controls while preserving existing permissions and business rules (Enterprise, Professional packages).
Vanta now exposes shared, permission-aware APIs for creating, viewing, updating, deactivating, and reactivating controls across Agent, MCP, and GraphQL surfaces (Professional package).
Mark audits complete and upload audit reports directly through the Vanta Agent and MCP.
Connect DocuSign directly to Vanta and ingest eligible contracts into Customer Commitments without manual uploads.
Connect Microsoft Copilot using workload identityworkload identityAn identity assigned to software, such as a service or container, rather than a person. It lets the workload authenticate and receive permissions without sharing a user's credentials. federation instead of a long-lived client secret.
Create internal TPRMTPRMAssessing 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. assessments that employees can complete without exposing the assessments to vendors (Enterprise package).
Link a risk scenario to multiple enterprise risks, with relationships visible across the Vanta Agent, REST APIREST APIA web API design style where clients interact with resources over HTTP using standard methods like GET, POST, and DELETE, giving cyber tools a predictable, language-agnostic way to integrate with external services., exports, and risk views.
Configure Custom Agents to proactively suggest updates to enterprise risk context when related child risks change.
Delete custom controls through the Vanta Agent or MCP, with associated mappings removed under the same authorization rules as the web app.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Through authorized MCP-connected agents, update common control fields, retrieve common control subdomain information, retrieve common control details and current status, retrieve deactivated control details without changing status (also via the Vanta Agent), and deactivate common controls while preserving existing permissions and business rules (Enterprise, Professional packages).
Vanta now exposes shared, permission-aware APIs for creating, viewing, updating, deactivating, and reactivating controls across Agent, MCP, and GraphQL surfaces (Professional package).
Mark audits complete and upload audit reports directly through the Vanta Agent and MCP.
Connect DocuSign directly to Vanta and ingest eligible contracts into Customer Commitments without manual uploads.
+3 moreshow less
Create internal TPRMTPRMAssessing 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. assessments that employees can complete without exposing the assessments to vendors (Enterprise package).
Link a risk scenario to multiple enterprise risks, with relationships visible across the Vanta Agent, REST APIREST APIA web API design style where clients interact with resources over HTTP using standard methods like GET, POST, and DELETE, giving cyber tools a predictable, language-agnostic way to integrate with external services., exports, and risk views.
Configure Custom Agents to proactively suggest updates to enterprise risk context when related child risks change.
CISO Assistant is a one-stop-shop GRC platform for Risk Management, AppSec, Compliance & Audit, TPRM, BIA, Privacy, and Reporting. It supports 200+ global frameworks with automatic control mapping, including ISO 27001, NIST CSF, SOC 2, CIS, PCI DSS, NIS2, DORA, GDPR, HIPAA, CMMC, and more.
CISO Assistant adds 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. findings tools, asset lifecycle tracking, EBIOS RMEBIOS RMA French risk-management method published by ANSSI for analysing cyber threats and selecting countermeasures; tools use it to structure risk assessments that satisfy French and EU regulatory expectations. study duplication, and new compliance frameworks including Swift CSCFSwift CSCFThe Customer Security Controls Framework published by SWIFT, defining mandatory and advisory cybersecurity controls that financial institutions must meet to connect to the SWIFT interbank messaging network. v2026/v2027 and SAMA BCM︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ GET THIS VERSION
$ git clone --branch v4.2.2 https://github.com/intuitem/ciso-assistant-community.git
# already have the repo? check out this version:$ git checkout v4.2.2
CISO Assistant adds MCP tools for findings assessment and writing findings, expanding the MCP surface with read and write operations on findings.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
CISO Assistant adds an instance time zone setting that controls timestamps in reports and exports.
CISO Assistant adds an asset lifecycle feature to track assets through defined lifecycle states.
CISO Assistant adds a new cumulative-based outcome model to support Essential Eight maturity logic.
CISO Assistant now lets workflows read requirement assignments and automatically promote their actors to respondents.
+5 moreshow less
CISO Assistant adds the ability to duplicate an EBIOS RM study.
CISO Assistant adds metrics creation in a single form with three-way direction, decimal display, and values-per-period tracking.
CISO Assistant adds Turkish PDF report templates.
CISO Assistant adds compliance library support for Swift CSCF v2026 and v2027, ASD Essentials (Oct. 2026), and the Australian Information Security Manual (ASD ISM, September 2026).
CISO Assistant adds the SAMA Business Continuity Management (BCM) Framework, the German legal text of the 2025 NIS-2 Implementation Act, and various enablers for the CyFun framework and Belgian market requirements to its compliance library.
CISO Assistant adds MCP tools for findings assessment and writing findings, expanding the MCP surface with read and write operations on findings.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
CISO Assistant adds an instance time zone setting that controls timestamps in reports and exports.
CISO Assistant adds an asset lifecycle feature to track assets through defined lifecycle states.
CISO Assistant adds a new cumulative-based outcome model to support Essential Eight maturity logic.
CISO Assistant adds the ability to duplicate an EBIOS RM study.
+2 moreshow less
CISO Assistant adds Turkish PDF report templates.
CISO Assistant adds the SAMA Business Continuity Management (BCM) Framework, the German legal text of the 2025 NIS-2 Implementation Act, and various enablers for the CyFun framework and Belgian market requirements to its compliance library.
Runlayer is an MCP gateway that enforces identity-mapped permissions and detects threats across the agents and MCP servers employees use.
Runlayer adds Artifacts (stable versioned links) and Apps (agent-built internal tools with owner-approved actions, a built-in database, scoped 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. access, and an audit log).
Runlayer now publishes agent-generated outputs as Artifacts — stable, versioned links requiring company sign-in that update in place when you ask the agent to revise them, with sharing controlled by people, groups, teams, or roles.
Runlayer now supports Apps — Artifacts with an owner-approved action list, a built-in database, and live access to business systems through existing MCP connectors, with no API keys stored in the app and each call running under the viewer's own permissions.
An Admin view lets app owners review every action before it runs, showing which MCP connectors it calls and whose access it uses; a changed action waits for re-approval while the last approved version continues to serve.
Every call an App makes is written to the Runlayer audit log alongside existing agent and MCP activity, recording which app, which action, which connector, and on whose behalf.
An AI-powered threat modeling tool that leverages OpenAI's GPT models to generate threat models for a given application based on the STRIDE methodology.
stride-gpt adds per-subsystem --max-llm-calls and --max-tool-calls caps, service-scoped agent planning, and a DFD editor in the web UI︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ GET THIS VERSIONpipxgit
pipx
$ pipx install stride-gpt==0.21.0
# already installed? upgrade to this version:$ pipx install --force stride-gpt==0.21.0
git
$ git clone --branch v0.21.0 https://github.com/mrwadams/stride-gpt.git
# already have the repo? check out this version:$ git checkout v0.21.0
PyPI attestation · checked 8 Nov
Caps 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. calls per subsystem with --max-llm-calls and tool calls per subsystem with --max-tool-calls, and plans subsystems by service while scoping each agent worker to its own files so analysis does not bleed across service boundaries.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Show a generated or parsed data flow diagram in the web UI editor after threat modelling.
Print the checkpoint path and resume command on Ctrl-C so interrupted runs can be restarted without digging through logs.
Improves LLM and tool call reliability by sending json_object instead of json_schema response format for DeepSeek models so structured calls succeed, and depending on litellm[mcp] so fresh installs can run tool calls without a separate 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. installation step.
Improves threat modelling by treating external services as trust-boundary components and covering hosted inference serving in the GenAI reference card, and reporting a threat chain that spans subsystems as an emergent cross-cutting threat rather than a simple aggregation.
+1 moreshow less
Add Sonnet 5.5 and Haiku 5.5 models; remove Sonnet 5, Haiku 4.5, and budget thinking.
Caps 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. calls per subsystem with --max-llm-calls and tool calls per subsystem with --max-tool-calls, and plans subsystems by service while scoping each agent worker to its own files so analysis does not bleed across service boundaries.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Show a generated or parsed data flow diagram in the web UI editor after threat modelling.
Print the checkpoint path and resume command on Ctrl-C so interrupted runs can be restarted without digging through logs.
Minimal is a secure, declarative runtimes for Claude Code, Gemini, Codex, Pi and more, so you can run multiple agents safely on your stack from a verified software supply chain.
minimal adds a Box Egress Proxy, secret management, 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. tracing, and a vulnerability reporting flag across CLI and daemon components︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ GET THIS VERSION
$ git clone --branch release-6a01d5ae https://github.com/gominimal/minimal.git
# already have the repo? check out this version:$ git checkout release-6a01d5ae
└──▷ TRY IT
Manage a secret token inside a box — store it once, then list to confirm it is registered.
$ min secret set MY_API_TOKEN
min secret list
Inspect a box's proxy-enforced audit trail and merge its parent chain for a full history.
$ min box audit --parent mybox
Submit a suspected vulnerability report from the local telemetry spool.
$ min bug --security
Adds min box audit to read a box's audit trail, min box audit --parent to merge parent chains, min doctor to report an edited audit chain, and min box spec to list each store reference with its authorities.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Report a suspected vulnerability to the project with --security on the diag command.
Sign in to GitHub and finish the OAuth flow on a real page with min login, with repository choice offered during sign-in.
Adds the bep crate implementing a Box Egress Proxy that issues leaf certificates under a name-constrained root and signing CA, holds proxy keys in the host Secure Enclave, maintains a hash-chained audit log per proxy decision, rotates the signing CA through a control socket, starts beside gvproxy so git and gh reach GitHub from inside VM boxes, and wires a client-key bound into bep serve via --max-client-keys to cap the registry size.
Name the QUICQUICA UDP-based transport protocol standardized by the IETF that bakes in TLS 1.3, multiplexed streams, and connection migration, giving cyber tools low-latency encrypted transport without head-of-line blocking. block helper in min finalize-install for post-install network configuration.
+4 moreshow less
Add opt-in OpenTelemetry export with local spooling across mlog, min, minimald, and minvmd, carrying one trace per request through CLI, daemon, and VM guest.
Give a project with no session entry an implicit default in mfile.
Keep box names answering across a reboot and a second VM.
Proxy runtime-published ports like a direct connection in minimald.
Adds min box audit to read a box's audit trail, min box audit --parent to merge parent chains, min doctor to report an edited audit chain, and min box spec to list each store reference with its authorities.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Report a suspected vulnerability to the project with --security on the diag command.
Sign in to GitHub and finish the OAuth flow on a real page with min login, with repository choice offered during sign-in.
Adds the bep crate implementing a Box Egress Proxy that issues leaf certificates under a name-constrained root and signing CA, holds proxy keys in the host Secure Enclave, maintains a hash-chained audit log per proxy decision, rotates the signing CA through a control socket, starts beside gvproxy so git and gh reach GitHub from inside VM boxes, and wires a client-key bound into bep serve via --max-client-keys to cap the registry size.
Add opt-in OpenTelemetry export with local spooling across mlog, min, minimald, and minvmd, carrying one trace per request through CLI, daemon, and VM guest.
+2 moreshow less
Give a project with no session entry an implicit default in mfile.
Proxy runtime-published ports like a direct connection in minimald.
2 more in Everything
8 more releases in this issue
· 2026-10-15 → 2026-11-07
minimal: Store, list, and remove secrets inside a box with min secret (subcommands: set, list, remove).︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ GET THIS VERSION
$ git clone --branch release-1e236070 https://github.com/gominimal/minimal.git
# already have the repo? check out this version:$ git checkout release-1e236070
└──▷ TRY IT
Store a GitHub token as a named secret so a box can retrieve it at runtime without it ever appearing in the toml.
$ min secret set GH_TOKEN
Inspect the full hash-chained audit trail for a box, merging in its parent chain, to verify no decisions were tampered with.
$ min box audit --parent mybox
Limit how many client keys the Box Egress Proxy will register before refusing new ones, to bound the registry surface.
$ bep serve --max-client-keys 50
Carry an edited-audit-chain warning into min doctor.
Carry telemetry spool and redact telemetry settings automatically when running min bug.
minimal: Control the client-key registry limit in bep serve with --max-client-keys.
└──▷ GET THIS VERSION
$ git clone --branch release-3c2b2a73 https://github.com/gominimal/minimal.git
# already have the repo? check out this version:$ git checkout release-3c2b2a73
└──▷ TRY IT
Store a secret in a box so an agent can access it without a manually pasted credential.
$ min secret set MY_API_KEY
Inspect a box's full audit trail, merging in its parent chain, to verify no decisions were tampered with.
$ min box audit --parent
Review every store reference a box declares along with the authorities bound to each, to audit egress policy before activating a session.
$ min box spec
Bundle the telemetry spool and redacted telemetry settings in min bug reports.
Sign in to GitHub through the published App and manage credentials with min login; logout deletes the sign-in and refuses every reference to it.
Ship a git credential helper inside grant boxes so agents can reach GitHub without a manually pasted PATPATA 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..
Block UDPUDPA network protocol that sends individual messages without guaranteeing delivery or order. Applications trade built-in reliability for less overhead and can add their own recovery. port 443 for boxes whose QUICQUICA UDP-based transport protocol standardized by the IETF that bakes in TLS 1.3, multiplexed streams, and connection migration, giving cyber tools low-latency encrypted transport without head-of-line blocking. stance is blocked, enforced both at the host and from a blocked cohort's own leaves.
Sign in to GitHub through the published App and manage credentials with min login; logout deletes the sign-in and refuses every reference to it.
Ship a git credential helper inside grant boxes so agents can reach GitHub without a manually pasted PATPATA 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..
Block UDPUDPA network protocol that sends individual messages without guaranteeing delivery or order. Applications trade built-in reliability for less overhead and can add their own recovery. port 443 for boxes whose QUICQUICA UDP-based transport protocol standardized by the IETF that bakes in TLS 1.3, multiplexed streams, and connection migration, giving cyber tools low-latency encrypted transport without head-of-line blocking. stance is blocked, enforced both at the host and from a blocked cohort's own leaves.
minimal: Introduces the new bep crate shipping the GitHub module and sealed-envelope proxy, with bep serve as the runtime entry point for egress control in the Box
└──▷ GET THIS VERSION
$ git clone --branch release-46d9e530 https://github.com/gominimal/minimal.git
# already have the repo? check out this version:$ git checkout release-46d9e530
└──▷ TRY IT
Store a GitHub token as a named secret in a box so agents can use it without it being pasted into the session environment.
$ min secret set GH_TOKEN
Review the tamper-evident audit trail for a box, merging its parent's history, to verify every proxy egress decision made during an agent session.
$ min box audit --parent <box-name>
Issue leaf certificates under a name-constrained root and signing CA; rotate the signing CA under an unchanged root, including rotation through the control socket.
Report non-Rust dependency pins and bump the locked_commit for packages in scripts.
Issue leaf certificates under a name-constrained root and signing CA; rotate the signing CA under an unchanged root, including rotation through the control socket.
minimal: Refuse Minimal's own infrastructure domains in the box CA and store rules.
└──▷ GET THIS VERSION
$ git clone --branch release-fad7c721 https://github.com/gominimal/minimal.git
# already have the repo? check out this version:$ git checkout release-fad7c721
└──▷ TRY IT
Store a secret token for a box so the agent can retrieve it without a manual paste.
$ min secret set MY_API_KEY
Inspect the tamper-evident audit trail for a specific box after a session ends.
$ min box audit <box-name>
List all store references and their declared authorities for a box to review egress policy.
$ min box spec <box-name>
Record a revocation when a live box is destroyed, and check minted box IDs against retained revocations.
minimal adds opt-in 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. tracing across the CLI, daemon, and VM host, plus a new bep proxy subsystem with certificate issuance, audit chaining, and egress enforcem︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ GET THIS VERSION
$ git clone --branch release-5378e56c https://github.com/gominimal/minimal.git
# already have the repo? check out this version:$ git checkout release-5378e56c
└──▷ TRY IT
Inspect the full proxy decision history for a specific box to verify egress enforcement.
$ min box audit
Adds the bep crate for proxied access control, including a GitHub module and sealed-envelope support, storing and publishing proxy key identities from the host keychain, hash-chained audit records for every proxy decision, leaf certificate issuance under a name-constrained root and signing CA, wiring redemption decisions and each box's declared egress into bep serve, refusing undeclared egress, host mismatches, off-port, off-set authorities, and expired values, and accepting only box connections while terminating TLS for module hosts over HTTP/1.1.
minimal adds opt-in 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. tracing across the CLI, daemon, and VM host, plus a signed macOS app bundle for minzoned︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ GET THIS VERSION
$ git clone --branch release-d67d78ee https://github.com/gominimal/minimal.git
# already have the repo? check out this version:$ git checkout release-d67d78ee
└──▷ HOW TO FIND IT
Opt out of OpenTelemetry telemetry export by setting the DO_NOT_TRACK environment variable before running any min command.
📍DO_NOT_TRACK=1 min session activate --attach .
minimal now exports OpenTelemetry spans from mlog with opt-in switches and a local spool, grounded in DO_NOT_TRACK opt-out support.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
minimal now traces each request end-to-end with one trace per request across both min and minimald.
minimal adds opt-in 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. export, min diag upload, min finalize-install, and staged-version installs via --version︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ GET THIS VERSION
$ git clone --branch release-6c6c39ae https://github.com/gominimal/minimal.git
# already have the repo? check out this version:$ git checkout release-6c6c39ae
└──▷ TRY IT
Install a specific staged release of minimal rather than the current stable channel.
$ curl --proto "=https" --tlsv1.2-fsSL 'https://go.minimal.dev/stable' | sh -s----version <version>
Run the new network finalisation step after install, reviewing the script before it applies.
$ min finalize-install --show--script
Install a specific staged version of minimal with --version on the installer.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Upload diagnostic bundles directly to the portal for analysis with min diag upload.
Replace min net setup with min finalize-install for host DNS and network setup; --show --script replaces --print.
Warn within a minute when the switch datapath is lost.
End a task run's box daemon-side.
+1 moreshow less
Give an absent egress destination list no network reach once layers merge.
6 details in Everything
└──▷ BREAKING ON UPGRADE
!The [harness] table in minimal.toml no longer loads a stack and now fails to load; rename it to [stack].
!--token is removed from min bug, min diag collect, and min diag upload; passing it is now a parse error. The Status: line is no longer printed after an upload.
!min net setup is removed with no alias; run min finalize-install instead (--show --script replaces --print).
!The own-ip deny-all egress default is now enforced for sessions and minvmd.
minimal adds per-session network isolation with egress rules, named addresses, port publishing, a root-setup command, multi-VM support, and CLI bug reporting︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ GET THIS VERSION
$ git clone --branch v0.7.0 https://github.com/gominimal/minimal.git
# already have the repo? check out this version:$ git checkout v0.7.0
└──▷ TRY IT
Activate a session that can only reach github.com and nothing else — useful for an agent that must pull code but must not exfiltrate data.
$ min session activate --network own_ip --allow-dns-hosts github.com --attach
Forward port 8080 from a running session to localhost:8080 so you can open a dev server in your browser without installing anything in the session.
$ min net forward mysession 8080:8080
Upload a diagnostic bundle immediately after a hang so the team can investigate without manual log collection.
$ min bug --upload
Adds session network control: min session activate --network <none|host_ip|own_ip> sets the network mode (host_ip shares the host network, own_ip gives the session its own IP, none removes all network access), --allow-dns-hosts <host>, --allow-subnets <cidr>, --deny-subnets <cidr>, --allow-protocols <tcp|udp|icmp>, and --deny-all-egress flags restrict what an own_ip session can reach, and min session policy <session> (also shown in the detail pane of min dash) inspects the effective network policy for a running session.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Adds session port management: --ingress <external>:<internal> publishes ports when a session starts, --dynamic-ingress <allow|ask|deny> and --dynamic-range <lo-hi> control runtime port requests, min net expose <port> requests a port from inside a session, and min net forward <session> <local>:<port> forwards a session port to localhost from outside.
Every session gets a <name>.min.internal hostname reachable via a local HTTP proxy, and via full DNS resolution after min finalize-install.
Collect and upload a diagnostic bundle in one step with min bug --upload, which prints a shareable report link and a delete command; upload a previously collected bundle with min diag upload <path>, and remove it later with min diag delete <id> <token> (bundles auto-expire after 7 days).
Collect all root-requiring setup into one reviewable sudo call with min finalize-install; preview what is missing with --show, print the exact script with --show --script, and undo the installation with --undo.
+1 moreshow less
Run multiple named VMs alongside the default with --vm <name> on min session activate.
6 details in Everything
└──▷ BREAKING ON UPGRADE
![harness] in minimal.toml no longer loads — rename the table to [stack].
!min ssh-forward is removed — use min net forward <session> <local>:<port> instead.
!The daemon's HTTPSHTTPSHTTP carried over an encrypted TLS connection. It protects traffic in transit and lets a browser verify the server's certificate. reverse proxy and client certificates are removed, and min login no longer issues a certificate — reach sessions via <session>.min.internal or min net forward.
!Session names must now be valid hostnames (letters, digits, hyphens only; case-insensitive; host, local, localhost, and minimald are reserved; names shaped like a session id are refused) — rename non-conforming sessions with min session rename.
BreachRx is an agentic AI CIRM platform that elevates cyber incident response to a governed enterprise-wide process.
BreachRx adds Slack and Microsoft Teams integrations to the Rex Platform, creating a private incident channel and direct-message action alerts per incident︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ TRY IT
Raise an incident from an ongoing Slack conversation without leaving the channel, so responders can begin formal tracking immediately.
$ /breachrx create
Connect Slack or Teams to the Rex Platform so private incident channels and action-owner DMs are created automatically for every new incident.
📍In the Rex Platform, go to Integrations, select Slack or Microsoft Teams, enable automatic channel creation, choose the roles to invite, and set notification rules for action assigned, updated, due soon, and overdue events.
BreachRx now creates a private Slack channel named breachrx-{identifier} for each new incident or exercise automatically, or on demand from the incident's menu.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Open a new incident from any Slack or Microsoft Teams channel using the /breachrx create slash command, which presents an intake form with incident name, type, severity, and description fields.
BreachRx adds a Microsoft Teams integration driven by the BreachRx Bot, replacing the older webhook-based connection, that creates private per-incident channels and auto-adds invited members to the team before the channel.
Direct messages now notify action owners in Slack when an action is assigned, updated, due soon, or overdue, with configurable notification rules.
Restricted incidents in Slack auto-create channels limited to the Admin role, the default channel member, and the incident creator; in Teams, restricted incidents invite only BreachRx organization administrators.
BreachRx now creates a private Slack channel named breachrx-{identifier} for each new incident or exercise automatically, or on demand from the incident's menu.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
BreachRx adds a Microsoft Teams integration driven by the BreachRx Bot, replacing the older webhook-based connection, that creates private per-incident channels and auto-adds invited members to the team before the channel.
Restricted incidents in Slack auto-create channels limited to the Admin role, the default channel member, and the incident creator; in Teams, restricted incidents invite only BreachRx organization administrators.
2 more in Everything
└──▷ BREAKING ON UPGRADE
!The Microsoft Teams integration replaces the older webhook-based connection with the BreachRx Bot; existing webhook-based Teams setups will need to be reconfigured.
Arcade is the actions runtime for enterprise AI agents â one control point to enforce, execute, and govern every action, every time.
Arcade adds Enterprise 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. Hosting so self-hosted customers can deploy custom MCP servers with the arcade deploy command︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
└──▷ TRY IT
Ship a custom MCP server to your self-hosted Arcade environment from a CI pipeline without filing tickets for a host or cluster access.
$ arcade deploy
Deploy custom MCP servers to self-hosted Arcade environments with the arcade deploy command from a local checkout or CI pipeline, without needing cluster access.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
Every deployed MCP server automatically runs behind Arcade's authorization, policy enforcement, and audit logging from its first call.
Operators can set deployment limits per organization and per project; when a team hits its limit, the error message states the limit plainly.
Monitor deployed servers through the dashboard's Servers page, which shows versioned releases, status, logs, and restart and removal controls.
Arcade adds Enterprise MCP Hosting support for Python servers built with the Arcade MCP framework, with FastMCP support planned next.
THE CYBER TOOLCHAIN