The daily firehose — everything the toolchain shipped today, already filtered.
We read every release from the 1309 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.
Open-source AI penetration testing tool to find and fix your app’s vulnerabilities.
Strix v1.1.0 adds 9 new security skills, SARIF 2.1.0SARIF 2.1.0A JSON-based file format, standardized by OASIS, for expressing static analysis results so that cyber tools from different vendors can share findings through a single, machine-readable schema. output, budget controls, and multi-target list scanning.
└──▷ GET THIS VERSION
$ git clone --branch v1.1.0 https://github.com/usestrix/strix.git
# already have the repo? check out this version:$ git checkout v1.1.0
└──▷ TRY IT
Scan a list of targets (e.g., all apps in scope for a bug-bounty engagement) from a file rather than passing each on the command line.
$ strix --target-list ./targets.txt
Adds weak-password detection and 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.prompt-injectionprompt-injectionAn attack where text the model reads, such as a web page, file or tool result, carries instructions that override the user's. It is the main security risk for agents with tools. security skills.
Adds GCP and Auth0 reconnaissance skills.
Adds five new security skills covering OAuth, AWS, prototype pollution, deserializationdeserializationTurning stored or transmitted data back into objects a program can use. Unsafe handling of untrusted data can allow unexpected behavior or code execution., and Django.
Adds skill-directory registration for extending the skill set.
Emits SARIFSARIFStatic Analysis Results Interchange Format, a standard for sharing code-analysis findings, including locations and rule details. It lets scanners report results to a common review interface. 2.1.0 output for GitHub code-scanning and ASPMASPMApplication security posture management, combining findings about applications from multiple tools and relating them to business context. It helps teams decide which fixes matter most. ingestion.
+10 moreshow less
Tags SARIF rules with STRIDE categories derived from CWECWECommon Weakness Enumeration, a catalog of recurring software and hardware weakness types. It describes classes of mistakes, such as injection, rather than individual vulnerability instances. mappings.
Adds --max-budget-usd flag to cap token/cost spend per scan.
Adds bind-mount option to handle large target repositories without copying them into the sandbox.
Adds --target-list CLI option to supply multiple targets from a file.
Adds new dependency reporting fields and requires advisory_cvss for dependency findings, with a new 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.TUITUIA text-based user interface that runs inside a terminal, using characters and ANSI escape codes to render menus, panes, and widgets without a graphical desktop, letting cyber tools ship rich interactivity with no GUI dependency. renderer.
Adds force_required_tool_choice setting for finer LLM tool-callingtool-callingA model's ability to answer with a structured request to run a named function or API, which the application executes and returns. It is how agents take actions. control.
Warns when the configured LLM is not a recommended or frontier model.
Routes ollama models through ollama_chat provider so tool calling works correctly.
Reports cost for streamed OpenRouterOpenRouterA unified API gateway that routes requests to dozens of LLM providers under one endpoint, letting cyber tools swap or compare models without rewriting provider-specific integration code. calls and attributes OpenRouter usage to the Strix app.
Adds TUI More content available hint for long views and throttled refresh with render caching to reduce scroll stutter.
BBOT 3.0 brings new HTTP/DNS engines, unified findings, 10+ new modules, and pydantic-validated config.
└──▷ GET THIS VERSION
$ git clone --branch v3.0.0 https://github.com/blacklanternsecurity/bbot.git
# already have the repo? check out this version:$ git checkout v3.0.0
└──▷ TRY IT
Regenerate a stale BBOT config file after upgrading to 3.0 to pick up new default keys.
$ bbot --reset-config
New in-process, rate-limited HTTP engine (blasthttp) replaces the external httpx subprocess for all web traffic.
New DNS resolution engine (BlastDNS) with bounded LRU caches and noisy PTR flood filtering.
Pydantic-validated config and preset system catches typos and type errors before a scan starts.
Scope reworked: -t/--targets defines scope and -s/--seeds drives the scan; ASNs now usable as scan targets.
Findings and Vulnerabilities unified into a single event type carrying Severity and Confidence.
+14 moreshow less
New --reset-config / --reset-secrets flags to regenerate stale config files.
New --no-color flag for color-free output.
Output modules are now additive; new exclude_output_modules option lets you opt individual modules out.
New output modules: MongoDB, Elasticsearch, Kafka, RabbitMQ, ZeroMQZeroMQA messaging library that provides socket-like abstractions for async message passing over TCP, IPC, or in-process transports, letting tools wire together distributed components without a dedicated message broker., and NATS.
New legba module for credential brute-forcing across many services.
New trajan module for misconfiguration scanning across GitHub, GitLab, Azure DevOps, Jenkins, and JFrog.
New virtualhost module for virtual-host fuzzing.
New waf_bypass module for WAFWAFA reverse proxy that inspects HTTP/HTTPS traffic and blocks requests matching known attack patterns such as SQL injection or XSS, sitting between clients and a web application to filter malicious input before it reaches the server.-bypass detection.
New shodan_enterprise module for Shodan Enterprise API integration.
New Hetzner object-storage bucket enumeration module.
Lightfuzz overhaul adds 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. and ESI submodules with a default WAF filter and reduced false positives.
HTTP wildcard/catch-all host detection cuts false positives on catch-all servers.
Findings de-duplicated in output and a severity breakdown shown in the status line.
Supported Python range expanded to 3.10–3.14.
19 details in Everything
└──▷ BREAKING ON UPGRADE
!-s now means --seeds (not --silent); scripts passing -s for quiet output will instead add a seed with no error. Silent output moved to -S/--silent.
!--whitelist was removed; use -t/--targets to define scope.
!--allow-deadly was removed.
!The noisy flag was renamed to loud (and other flags were similarly renamed).
!Module options are now pydantic-validated; the legacy per-module options dict hard-fails on upgrade.
!Some event types, event attributes, and config keys were renamed or removed.
!Modules wappalyzer, smuggler, digitorus, sitedossier, passivetotal, and wpscan were removed.
Portkey is an AI gateway and observability platform that helps developers manage, monitor, and optimize LLM API calls with routing, caching, and reliability features.
└──▷ USE IT
Gate every agentic request through Zscaler AI Guard so your existing Zscaler Detection Policies apply to 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. inputs and outputs.
json
In your Portkey config, add the Zscaler AI Guard guardrail with your policyId:
{
"guardrails": {
"beforeRequestHook": {
"provider": "zscaler",
"policyId": "<your-zscaler-policy-id>",
"timeout": 10000
},
"afterRequestHook": {
"provider": "zscaler",
"policyId": "<your-zscaler-policy-id>",
"timeout": 10000
}
}
}
Route GCP Cloud Storage log writes from an AWS-hosted gateway through Workload Identity Federation to avoid storing long-lived GCP service account keys in AWS.
$ Set these environment variables on your gateway deployment:
GCP_WIF_AUDIENCE=<your-wif-audience>
GCP_WIF_SERVICE_ACCOUNT_EMAIL=<your-sa>@<project>.iam.gserviceaccount.com
Adds Secret References so Portkey fetches credentials at runtime from AWS Secrets Manager, Azure Key Vault, or HashiCorp Vault via secret_mappings, keeping keys out of Portkey config entirely.
Enables GCP Workload Identity Federation log delivery from AWS-hosted gateways to Google Cloud Storage via GCP_WIF_AUDIENCE and GCP_WIF_SERVICE_ACCOUNT_EMAIL, eliminating long-lived GCP keys in AWS.
Adds customHost support on the Bedrock Guardrails plugin so guardrail evaluation hits private or regional Bedrock-compatible endpoints instead of default public URLs.
Adds tool_choice, stream_options, and streaming usage reporting to DeepSeek deepseek-chat; maps reasoning_effort to thinking mode and surfaces reasoning_content in streams for deepseek-reasoner.
Maps Portkey metadata to Vertex AI resource labels and adds enterprise search grounding via enterpriseWebSearch / enterprise_web_search with separate cost attribution.
+2 moreshow less
Enables AWS workloads to reach Vertex AI through AWS–GCP WIFWIFWorkload Identity Federation is a Google Cloud mechanism that lets external workloads authenticate to Google APIs using short-lived tokens from a trusted identity provider, eliminating the need to manage long-lived service account keys. using GCP_WIF_AUDIENCE and GCP_WIF_SERVICE_ACCOUNT_EMAIL.
Updates OpenTelemetryOpenTelemetryAn open standard and set of tools for collecting traces, metrics and logs from applications. It lets services export observability data without tying instrumentation to one monitoring vendor. GenAI semantic spans to semconv 1.40.0, adds OTELOTELAn 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. exporter support for guardrail flows, and supports custom resource attributes for downstream APMAPMApplication Performance Monitoring — a category of software that collects metrics, traces, and logs from running applications so developers can detect slowdowns and diagnose their root cause. standardization.
Hunt every Endpoint in your code, expose Shadow APIs, map the Attack Surface.
Noir v1.2.0 adds CLI app surface detection, 9 stdlib HTTP servers, schema-driven APIs, enterprise legacy stacks, and major perf gains.
└──▷ GET THIS VERSION
$ git clone --branch v1.2.0 https://github.com/owasp-noir/noir.git
# already have the repo? check out this version:$ git checkout v1.2.0
Detects CLI application attack surfacesattack surfacesThe exposed systems, interfaces and access points an attacker could try to use. It includes internet-facing services as well as identities, software and other reachable resources. (argv, flags, env vars) across 17 languages and 40+ libraries including Cobra, Click, clap, Commander, Thor, and picocli.
Adds built-in HTTP server analyzers for Go (net/http), Python (http.server), Node (node:http), Java (com.sun.net.httpserver), C# (HttpListener), Ruby (WEBrick), Crystal, Dart, and Zig.
Adds schema-generated API analyzers for Strapi, Directus, Payload CMSCMSA web content management system for creating and managing digital content without writing code directly; cyber tools use CMS platforms to host documentation, portals, or reporting dashboards quickly., Supabase/PostgREST, Hasura, and Appwrite — extracting routes and params from schema files rather than source routes.
Adds enterprise legacy stack support: CFML/ColdFusion, Classic ASP, ASP.NET WebForms, and CFML frameworks Taffy, ColdBox, Wheels, and FW/1.
Adds BEAM language support: Erlang (Cowboy, Elli) and Gleam (Wisp).
+9 moreshow less
Adds PHP CMS/platform analyzers for WordPress (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., admin-ajax, admin-post), Drupal (*.routing.yml), and Magento (Web API + MVC controller routes).
Adds Django Ninja (Python) and Plumber (R) framework analyzers.
Adds Terraform/OpenTofu analyzer to extract AWS API Gateway v1/v2 routes from .tf HCL and .tf.json files, resolved module-wide.
Adds realtime/event endpoint detection for SignalR, Socket.IO, Phoenix Channels, and Action Cable.
DataDome is a bot protection and fraud prevention platform that detects and blocks malicious traffic, scraping, and cyberattacks in real time.
DataDome Fastly Compute 1.4.0 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. protocol support with HTTP message signature collection per RFC 9421RFC 9421An IETF standard defining HTTP Message Signatures, a scheme for cryptographically signing HTTP requests and responses so tools can verify message integrity and authenticity end-to-end..
└──▷ USE IT
Enable MCP traffic inspection by pointing DataDome at your MCP endpoint in the Fastly Compute service config.
ini
datadome_mcp_endpoint = "/mcp"
Adds Model Context ProtocolModel Context ProtocolModel 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. (MCP) support with new datadome_mcp_endpoint configuration parameter (defaults to /mcp) to specify the MCP endpoint.
Collects Signature, Signature-Agent, and Signature-Input headers to support HTTP message signatures as defined by RFC 9421.
Collects Skyfire-Pay-ID header from incoming requests.
iLEAPP v2026.2.0 adds Biome, C2PA AI provenanceprovenanceEvidence recording where a software artifact came from and how it was built, including its source and build process. It helps a consumer check that the artifact matches a trusted build., and United Airlines artifacts with full iOS 26 and Python 3.13/3.14 support.
└──▷ GET THIS VERSION
$ git clone --branch v2026.2.0 https://github.com/abrignoni/iLEAPP.git
# already have the repo? check out this version:$ git checkout v2026.2.0
└──▷ TRY IT
Parse an iOS 26 extraction to capture new Biome and C2PA artifacts in a full HTML/TSVTSVA plain-text format storing tabular data with fields separated by tab characters, making it easy for cyber tools to parse and exchange structured records without a CSV parser's quoting complexity. report.
$ ileapp -t zip -i /path/to/ios26_extraction.zip -o /path/to/output/
Limit a parse run to only the new Biome and C2PA modules using a saved profile, to reduce processing time on large extractions.
Adds 14+ new Biome database artifacts parsing SQLite stores under Library/Biome/databases/ (app openings, cable plug events, backlight events, CarPlay activity, intelligence platform entities, and more).
Adds Biome Set.db artifacts for installed apps and other syncable streams under Library/Biome/sets/.
Adds Biome stream artifacts for harvested mail/messages, reads, screen time, and keyboard data.
Adds C2PA Content Provenance artifact parsing AI-generated image provenance, including signer certificate, signing time, Media State column, and case-insensitive media path matching.
Adds new United Airlines app artifact module extracting travel data from the United Airlines iOS app.
+3 moreshow less
Reworks Photos.sqlite parser with full iOS 26 support, context form, and generator compatibility.
Adds Python 3.13 and 3.14 support.
Expands iOS test corpus coverage to 13 archives spanning iOS 13.3.1 through the latest releases.
HyperDbg v0.21 adds Intel PT parameter parsing, TID/PID helpers, hyperperf infrastructure, and an SDK load-state API.
└──▷ GET THIS VERSION
$ git clone --branch v0.21 https://github.com/HyperDbg/HyperDbg.git
# already have the repo? check out this version:$ git checkout v0.21
Adds user-mode and kernel-mode Intel PT parameter parser for fine-grained control of processor trace configuration.
Adds thread (TID) and process (PID) helper functions for Intel PT, enabling per-thread and per-process tracing workflows.
Introduces the hyperperf (Hardware Performance Counter) project structure as a new HyperDbg SDK module.
Adds hypertrace, hyperevade, and hyperperf DLLDLLA Windows file format containing compiled code and data that multiple programs can load and share at runtime, letting cyber tools inject functionality into other processes or hook system calls. files to the SDK, expanding the available module surface.
Exports a new SDK API for checking whether any HyperDbg module is currently loaded.
+2 moreshow less
The unload command now reports whether any module is loaded before attempting to unload, improving operator feedback.
Separates SDK libraries into distinct user-mode and kernel-mode modules for cleaner integration.
Vectra AI is a cybersecurity platform that uses AI-driven threat detection and response to identify and prioritize attacks across hybrid cloud, network, and identity environments.
Adds behavioral signatures for HTTP-based attacker tooling, including malicious file downloads/uploads and certificate-related abuse via trusted Windows utilities and offensive frameworks.
Sophos Central is a cloud-based management platform that unifies administration of Sophos security products, including endpoint, network, email, and cloud protection.
Sophos Central adds Network Map, Live Discover for AP6, DNS Protection APIs, Protected Browser extension, and a wave of new alerts and email capabilities.
└──▷ HOW TO FIND IT
View the Network Map to understand firewall-to-switch-to-AP6 topology and spot unexpected device relationships at a glance.
📍In Sophos Central, go to My Products › Network › Network Map.
New Network Map provides a visual topology of all Sophos firewalls, switches, and AP6 access points with search and relationship views.
Live Discover queries now support AP6 Series access points, enabling on-demand SQL queries against wireless data in the Sophos data lake.
New Protected Browser extension lets admins apply security policies and monitor user activity in Chrome and Edge without requiring migration to the full Protected Browser.
Web control adds filtering for Generative AI sites, more granular per-site controls, and simplified policy management.
New alerts for Account Health Check (AHC) issues now surface unaddressed AHC findings directly in the alerts feed.
+29 moreshow less
Protected Browser gains a categorized app picker for Generative AI policy creation and new reporting on file uploads and downloads.
DNS Protection endpoint and DNS filtering policies are now managed on a single page for Workspace Protection customers.
Administrators will soon be required to register at least two MFAMFAMulti-factor authentication, requiring proof from different categories, such as a password and a hardware key, to sign in. It reduces reliance on a password alone. methods; a grace period allows registration of a second method at sign-in.
New alert fires when API credentials (including AD Sync) are 30, 7, or 1 day from expiration.
New alert fires when a Super Admin or Admin role is assigned or removed.
Public RESTful APIs for DNS Protection core configuration are now generally available.
New Security Services Retainer provides flexible access to Advisory and Professional services plus guaranteed emergency incident response.
Global Settings replaces General Settings with use-case-driven categories, search, filters, and a left-panel index.
New Sophos Email Plus license tier extends Message History and Quarantine retention to 90 days, adds Message Replay, Contact Trace and Recall, and bundles DMARC Manager.
M365-quarantined emails can now be viewed and managed directly within Sophos Central without switching to the M365 portal.
Email: Newly Registered Domain (NRDNRDNewly Registered Domain — a domain name created within a recent window (often 30–90 days); cyber tools flag these because attackers frequently register fresh domains to evade reputation-based blocklists.) Protection detects and acts on emails from recently registered domains before delivery.
Sophos Email TLS report now includes MFR-mode emails, providing full visibility into TLS encryption across all processed mail.
Workspace Protection is generally available, bundling Sophos Protected Browser, DNS Protection, ZTNAZTNAZero Trust Network Access, a framework that grants users access to specific applications only after continuously verifying identity and device posture, replacing broad VPN-style network access with least-privilege connections., and Email Monitoring System.
Device tagging is generally available with tag-based filtering, search-box tag queries, configurable tag column, install-time tagging, and tag suggestions.
Partner Super Admins can now reset MFA for users and admins in managed customer tenants.
PSA tickets for Autotask and ConnectWise are now enriched with Integrated MSPMSPA managed service provider, a company that operates IT services for customers. It may manage systems, networks, backups or other ongoing technology needs. Tool data.
AI Assistant adds access to Self-managed and Sophos-managed cases, pinned content, structured feedback tags, and the ability to save and reuse custom prompts.
Unauthorized File Protection is a new capability that prevents malicious executables from running on Windows servers.
AP6 Series MR8 adds WPA2/WPA3WPA3The third generation of the Wi-Fi Protected Access authentication and encryption standard, maintained by the Wi-Fi Alliance, offering stronger handshake protection and per-session encryption compared to WPA2. Enterprise Authentication, adjustable MAC filtering, and increases the maximum MAC list size to 4096 addresses per SSID.
Managed Risk now includes authenticated internal vulnerability scanning for deeper visibility beyond unauthenticated scans.
Sophos Switch management gains STPSTPSpanning Tree Protocol, an IEEE standard (802.1D) that prevents Layer 2 network loops by electing a root bridge and blocking redundant switch paths, keeping Ethernet networks stable under failure conditions. (Spanning Tree Protocol) configuration in Sophos Central.
Email TLS report shows which TLS version was applied per message and reasons for non-delivery due to Secure Message Policy TLS requirements.
AI Search for natural-language queries against security data is generally available for all Sophos XDRXDRExtended detection and response, combining signals and investigation across endpoints, networks, identities or cloud services. It helps analysts connect activity that separate tools might miss. and 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. customers, now including endpoint telemetry.
AI Case Summary and AI Command Line analysis are generally available for accelerating investigations.
Sophos DMARC Manager add-on provides DMARC, SPFSPFAn email authentication standard that lets a domain owner publish, via DNS, which mail servers are authorized to send on its behalf, so receiving servers can reject spoofed senders. (with flattening), and DKIMDKIMAn email authentication standard that lets a sending mail server cryptographically sign outgoing messages so receiving servers can verify the signature was not forged or tampered with in transit. management and reporting, plus BIMI, MTA-STSSTSAWS Security Token Service, an Amazon Web Services API that issues short-lived credentials for IAM roles, letting cyber tools authenticate to AWS without storing long-term secrets., and TLS-RPT.
Integration health alerts now fire in Sophos Central and via email when a third-party MDR/XDR data ingestion integration is unhealthy or offline.
Central Switch Stacking provides a management overlay to simplify administration of stacked Sophos switches.
Usage Insights for AP6 Series access points is now available in Sophos Central Wireless.
Trend Vision One is a unified cybersecurity platform that integrates extended detection and response (XDR) across endpoints, networks, email, and cloud environments.
Trend Vision One extends max inactivity and session timeout limits from 8 hours to 12 hours.
└──▷ HOW TO FIND IT
Extend session limits for SOCSOCA security operations center, the team or function that monitors threats, investigates alerts and coordinates incident response. Its work can be internal, outsourced or shared. analysts running multi-hour threat investigations without re-authentication prompts.
📍In the console, go to Administration → Console Settings, open the Timeout section, and set Max session timeout and Inactivity timeout to up to 12 hours.
Raises the configurable ceiling for Max session timeout from 8 hours to 12 hours, allowing administrators to keep analysts signed in longer during extended investigations.
Raises the configurable ceiling for Inactivity timeout from 8 hours to 12 hours (capped at the Max session timeout value), reducing re-authentication interruptions during complex tasks.
Laurel v0.8.0 adds SHA-256SHA-256A cryptographic hash function that turns data into a fixed 256-bit fingerprint. It is used to check integrity; hashing does not encrypt the original data. exec hashing, UNIX socket logging, env-var prefix matching, and unified process spawning tracking.
└──▷ GET THIS VERSION
$ git clone --branch v0.8.0 https://github.com/threathunters-io/laurel.git
# already have the repo? check out this version:$ git checkout v0.8.0
New SYSCALL.SPAWNED_BY field unifies process-lineage tracking across both single-exec and fork/exec scenarios.
Supports prefix matching for environment variables, enabling selective capture of env vars by prefix.
Adds logging output to a UNIX domain socket for integration with local log consumers.
Supports calculating SHA-256 hashes for executed binaries, enabling integrity checking of process execution.
OpenCTI 7.260715.0 adds connector log TTLTTLTime to live, a limit on how long data remains valid or how far a packet can travel. Caches use a time limit; IP packets use a hop limit. and ingestion history in Redis, plus authentication validity policy improvements.
└──▷ GET THIS VERSION
$ git clone --branch 7.260715.0 https://github.com/OpenCTI-Platform/opencti.git
# already have the repo? check out this version:$ git checkout 7.260715.0
Adds connector logs TTL and ingestion history storage in Redis for improved connector observability.
Improves the authentication validity policy update workflow.
2 details in Everything
6 more releases in this issue
· 2026-06-24 → 2026-07-15
OpenCTI 7.260710.0 adds admin-forced password resets, Custom Views date ranges & attribute widgets, OAuth 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. dynamic tokens, and security coverage custom views.
└──▷ GET THIS VERSION
$ git clone --branch 7.260710.0 https://github.com/OpenCTI-Platform/opencti.git
# already have the repo? check out this version:$ git checkout 7.260710.0
└──▷ HOW TO FIND IT
Narrow a Custom View on an entity page to a specific investigation window without changing global filters.
📍In the console, open any entity page, navigate to a Custom View tab, and use the new date range picker in the view configuration to set a start and end date.
Force a user to reset their password at next login for account hygiene or after a suspected compromise.
📍In the console, go to Settings › Security › Users, select the target user, and use the new Force password change action.
Enables admins to force a password change for users via a new authentication capability.
Adds date range configuration directly within Custom Views on entity pages.
Adds an attribute widget for Custom Views.
Extends Custom Views to the security coverage section.
Supports dynamic token (OAuth) for SMTP email configuration.
+3 moreshow less
Renames connector catalog statuses from Verified/'Unverified' to Supported by Filigran/'Supported by Community'.
Adds telemetry for shared saved filters.
Adds search-engine crawler blocking to prevent web-site indexing of the platform.
OpenCTI 7.260309.0-lts.6 improves automatic enrichment on element creation/update and enhances JSON mapper regex extraction.
└──▷ GET THIS VERSION
$ git clone --branch 7.260309.0-lts.6 https://github.com/OpenCTI-Platform/opencti.git
# already have the repo? check out this version:$ git checkout 7.260309.0-lts.6
Improves automatic enrichment triggered on creation or update of elements.
Enhances JSON mapper extraction with improved regexp support.
OpenCTI 7.260706.0 ships a native MCPMCPModel Context Protocol, a standard for connecting AI applications to external tools and data sources. A compatible server exposes capabilities an assistant can use through a common interface. server, STIXSTIXStructured Threat Information Expression, a standard format for describing threats, indicators and their relationships. It helps organizations exchange threat intelligence in a usable form. security-coverage export, custom newsfeed views, and richer telemetry.
└──▷ GET THIS VERSION
$ git clone --branch 7.260706.0 https://github.com/OpenCTI-Platform/opencti.git
# already have the repo? check out this version:$ git checkout 7.260706.0
└──▷ HOW TO FIND IT
Find your MCP server connection details to integrate OpenCTI with AI assistants via XTM One.
📍In the OpenCTI console, go to your User Profile and locate the XTM One MCP server section to retrieve the Streamable HTTP endpoint URL and credentials.
Adds a native Model Context ProtocolModel Context ProtocolModel 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. (MCP) server with Streamable HTTPStreamable HTTPAn MCP transport protocol that lets a client send JSON-RPC requests over HTTP POST while the server streams responses back using Server-Sent Events, replacing the older stdio and SSE transports for remote tool communication. transport, exposed through XTM One and surfaced in the user profile.
Enables export of security coverage data to STIX format.
Supports custom views as a resource type in the newsfeed.
Adds TELEMETRY_TAGS deployment dimensions and tracks active connectors by catalog identity.
Extends telemetry to cover full AI usage and product adoption metrics, including workflow publish counts.
+5 moreshow less
Maintains telemetry visibility for managed connectors absent from the catalog using a registry-stripped image identity.
Introduces a data sanity manager as an alternative to certain database migrations.
Improves JSON mapper extraction with enhanced regexp support.
Enhances notification access control for workflow recipients.
Improves the engine retry mechanism for greater resilience.
OpenCTI 7.260701.0 adds XTM Hub custom view deployment, stream URI deny lists, ES health monitoring, and workflow transition comments.
└──▷ GET THIS VERSION
$ git clone --branch 7.260701.0 https://github.com/OpenCTI-Platform/opencti.git
# already have the repo? check out this version:$ git checkout 7.260701.0
Enables deploying custom views received from XTM Hub directly into OpenCTI.
Adds a configurable deny list of URIs to block during stream ingestion.
Adds an Elasticsearch engine health monitor to surface ES issues proactively.
Supports adding comments on workflow transitions and viewing them in the UI, with notifications sent to assignees and participants.
OpenCTI 7.260626.0 adds draft filtering in dashboards, a new text editor, and tightened AI playbook agent controls.
└──▷ GET THIS VERSION
$ git clone --branch 7.260626.0 https://github.com/OpenCTI-Platform/opencti.git
# already have the repo? check out this version:$ git checkout 7.260626.0
└──▷ HOW TO FIND IT
Track draft workspace activity across your team by adding a dashboard widget filtered to draft entities.
📍In the console, go to Dashboards › (your dashboard) › Add Widget, then in the entity filter select Draft Workspace as the entity type to scope the visualization to draft data.
Limit which identities an AI playbook component can impersonate when running automated enrichment, reducing privilege-escalation risk.
📍In the console, go to Automation › Playbooks › (your playbook) › AI Agent component › Run as field, and select from the restricted list of the current user or available service accounts.
Adds drafts as filterable entities in dashboard widgets, enabling visualizations and metrics over draft workspaces.
Restricts the AI agent run as selection in playbooks to the current user and service accounts only, tightening execution scope.
Replaces the legacy text editor with a new editor solution across the platform.
Enables entity lookup with on-the-fly creation in form intakes for users who have only create/update-in-draft permissions, without triggering a full creation process.
OpenCTI 7.260624.0 adds bulk playbook enrollment, has/not-changed stream filters, CTEMCTEMContinuous threat exposure management, a recurring process for finding, prioritizing and validating exploitable exposures. It connects discovery to fixes and checks whether those fixes work. shortcut, and trust-proxy IP whitelisting config.
└──▷ GET THIS VERSION
$ git clone --branch 7.260624.0 https://github.com/OpenCTI-Platform/opencti.git
# already have the repo? check out this version:$ git checkout 7.260624.0
└──▷ HOW TO FIND IT
Use the has/not-changed operator in a playbook filter to trigger actions only when a specific attribute on an entity has changed since the last event.
📍In the Playbook or Stream filter editor, add a filter condition, select the desired attribute, and choose the has changed or not changed operator from the operator dropdown.
Bulk-enroll a large set of existing entities into a playbook using smart filtering instead of enrolling one at a time.
📍In the Playbooks UI, open a playbook, click Bulk enrollment, apply the desired entity filters, and confirm to enroll all matching entities at once.
Adds trust proxy configuration for IP whitelisting, enabling accurate client IP resolution when OpenCTI sits behind a reverse proxy.
Adds a CTEM Command Center shortcut to the top bar when XTM One is connected.
Adds a feature flag to revert to the old rich text editor alongside the new @filigran/rich-text-editor component.
Adds dynamic user resolution in the playbook send to notifier component, allowing recipients to be resolved at runtime.
Adds has/not-changed attribute filtering operators for playbook and stream filtering components.
+3 moreshow less
Enables bulk enrollment of entities into a playbook with smart filtering.
Streamlines Autonomous System creation by auto-populating the AS number from the Autonomous-System name during ingestion.
Renames validated draft's status label to Validation progress in the draft workflow UI.
Watcher - Open Source AI-powered Cyber Threat Intelligence & Hunting Platform. Developed with Django & React JS.
Watcher v3.5.0 adds a Timeline audit trail, UDRP approval workflow, a reworked ThreatsWatcher dashboard, and auto-fit table pagination.
└──▷ GET THIS VERSION
$ git clone --branch v3.5.0 https://github.com/thalesgroup-cert/Watcher.git
# already have the repo? check out this version:$ git checkout v3.5.0
└──▷ HOW TO FIND IT
Review and approve a queued UDRP domain transfer so no won case promotes a domain without human sign-off.
📍In the Watcher UI, navigate to the PendingActions list, locate the entry with status pending, review the payload, then click Approve (or Reject) to trigger or cancel the domain transfer.
Adds a full Timeline audit trail module that records every create, update, and delete action across all Watcher modules, with per-object history panels, field-level diffs, and actor avatars visible in every table row.
New /api/timeline/events/ endpoint lets practitioners query the full history of any tracked object by content type and object ID.
Introduces a PendingActions approval workflow for UDRP case transfers: when a case is marked won, a PendingAction is queued for admin review instead of automatically calling transfer_to_legitimate_domains().
Adds API endpoints for listing, approving, and rejecting pending actions, secured by existing permission guards.
New avatar_color field on UserProfile automatically assigns a deterministic color to each user on first login, surfaced as color-coded initials throughout the UI.
+2 moreshow less
ThreatsWatcher dashboard redesigned with a new ResizableContainer split-pane layout, a WordSummary AI summary panel alongside post URLs, and a draggable divider with localStorage persistence.
Removes Google Fonts CDNCDNA content delivery network, servers distributed across locations that cache and serve content near readers. It can reduce loading time and traffic to the original server. calls; all icon and font assets are now served locally, eliminating external network dependencies at page load.
7 details in Everything
└──▷ BREAKING ON UPGRADE
!The new timeline Django app requires a database migration (python manage.py migrate) and must be present in INSTALLED_APPS with its URLs registered — pull the updated settings.py and urls.py or add them manually before starting the service.
OpenCVE 3.0.0 ships Automations — trigger-driven alert and report workflows with conditions, actions, and webhooks.
└──▷ GET THIS VERSION
$ git clone --branch v3.0.0 https://github.com/opencve/opencve.git
# already have the repo? check out this version:$ git checkout v3.0.0
└──▷ USE IT
Disable public self-registration on a self-hosted instance to lock down new account creation.
yaml
ENABLE_REGISTER: false
New Automations engine lets teams define CVE workflows with triggers, conditions, and actions (notifications, assignments, status changes, webhooks).
Alert automations react to matching CVEs and execute configurable actions including notifications, assignments, status changes, or webhooks.
Report automations deliver daily or weekly CVE digests with configurable conditions and timezone support.
Default report and alert automations are created automatically during project onboarding.
Reports now include CVE and 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. summaries plus weekly AI-generated summarization.
+14 moreshow less
Enhanced reports dashboard widget for at-a-glance digest tracking.
Filters added to notifications and reports listings.
Test notification action and notification type picker modal for verifying delivery channels.
Redesigned organization pages with tabs, member role management, and audit logs.
Comments now supported in CVE tracking.
OpenCVE Enrichment versioning added to CVE details.
Project vulnerabilities can be filtered by multiple statuses simultaneously, or by vulnerabilities without any status.
NVDNVDThe National Vulnerability Database, maintained by NIST, is the U.S. government repository of CVE records enriched with CVSS scores and metadata; cyber tools query it to drive vulnerability prioritization.CPECPECommon Platform Enumeration, an NIST-maintained naming scheme that gives software, hardware, and operating systems a structured, machine-readable identifier so cyber tools can match vulnerabilities to affected products consistently. configurations now exposed in the CVE API response.
New ENABLE_REGISTER setting allows administrators to disable new user signups.
Administrators can now reset user MFAMFAMulti-factor authentication, requiring proof from different categories, such as a password and a hardware key, to sign in. It reduces reliance on a password alone..
CVSS version selection added to the onboarding flow.
MITREMITREA US federally funded research and development center that maintains public frameworks—ATT&CK, CVE, D3FEND—that cyber tools use to classify threats, vulnerabilities, and defensive techniques in a common language. weaknesses synced with improved browsing and search.
Vendors load incrementally without a full page reload.
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. blocked in webhook and Slack notification channels.
Immuta is a data security platform that automates access control, privacy policy enforcement, and compliance monitoring across cloud data environments.
Immuta adds managed groups for agentic policies, Redshift viewless discovery, Delta Sharing enforcement, and OAuth 2.0OAuth 2.0An open authorization framework (IETF RFC 6749) that lets a service grant limited API access to a third party without sharing credentials, giving cyber tools a standard way to authenticate against external platforms. for Alation.
└──▷ HOW TO FIND IT
Restrict AI-sensitive data sources so only human (non-agent) accounts can subscribe, using the new built-in All Agents group.
📍In the Immuta UI, go to Policies › Subscription Policies › New Policy. Under Who is denied, select the built-in group All Agents. Apply the policy to data sources tagged as AI-restricted.
Switch Alation integration to OAuth 2OAuth 2An open authorization framework (RFC 6749) that lets a service grant limited access to a user's resources on another service without exposing credentials, giving cyber tools a standard way to authenticate against third-party APIs..0 to eliminate 90-day refresh-token rotation for the Immuta–Alation connection.
📍In the Immuta UI, go to App Settings › Integrations › Alation, select OAuth 2.0 as the authentication method, and supply your OAuth 2.0 client credentials.
Enable automated sensitive data discovery on a Redshift viewless data source to auto-tag PIIPIIPersonally Identifiable Information — data that can identify a specific individual, such as names, email addresses, or SSNs. Cyber tools flag or redact it to limit exposure and meet privacy regulations. columns.
📍In the Immuta UI, open your Amazon Redshift viewless data source, navigate to the Sensitive Data Discovery settings, and enable automated discovery. Immuta will scan columns using regex, dictionary, and column-name-regex identifiers and apply matching tags automatically.
Adds three built-in system-managed groups (All Agents, All Users, All Identities) usable in subscription and data policies to enforce system-wide access rules without manual group management.
Supports row filtering and column masking on Unity Catalog managed, external, and foreign tables in Databricks Delta Sharing shares.
Adds OAuth 2.0 authentication option for Immuta's Alation integration, eliminating recurring API key rotation requirements.
Extends Databricks Unity Catalog tag ingestion to foreign catalogs (Lakehouse FederationLakehouse FederationA data architecture pattern that lets a single query engine read from multiple external data sources—databases, data lakes, warehouses—without moving or copying the data, reducing pipeline complexity for tools that need broad data access.) by falling back to system.information_schema tag views.
Embeds Immuta documentation and an AI-powered chatbot directly in the Immuta UI, accessible from the top menu Documentation icon.
+1 moreshow less
Introduces case-insensitive Starburst (Trino) principal name matching when mapping Trino principals to Immuta users and service accounts.
6 details in Everything
└──▷ BREAKING ON UPGRADE
!Snowflake policy role names will be renamed as part of the back-end architecture update rolling out beginning July 23, 2026; existing policy roles will be recreated under new names (user role names remain unchanged).
!Integration error banners for Snowflake integration validation failures will no longer be displayed in the existing location; errors are now only available through the new connections experience.
!When a data source is disabled in Immuta, data owners' Immuta-managed grants in the data platform will be revoked (behavior change release: on by default 8/1, enabled for all 9/1).
authentik 2026.5.5 adds number and boolean support in AKQL queries on JSON fields and licensing usage totals in the admin UI.
└──▷ GET THIS VERSION
$ git clone --branch version/2026.5.5 https://github.com/goauthentik/authentik.git
# already have the repo? check out this version:$ git checkout version/2026.5.5
Adds number and boolean type support in AKQL queries on JSON fields via the API search interface.
Adds usage totals display to the licensing section in the admin web UI.
2 details in Everything
1 more release in this issue
· 2026-07-07 → 2026-07-15
authentik 2026.2.5 adds SCIMSCIMA standard for creating, updating and removing user accounts and groups across applications. It helps an identity provider keep access in sync when people join, change roles or leave. discovery support and in-process per-IP rate throttling.
└──▷ GET THIS VERSION
$ git clone --branch version/2026.2.5 https://github.com/goauthentik/authentik.git
# already have the repo? check out this version:$ git checkout version/2026.2.5
Adds discovery support to the SCIM provider, enabling automatic capability negotiation with SCIM endpoints.
Introduces in-process per-IP rate throttling to limit abusive request rates without an external component.
RegScale is a compliance operations platform that automates and continuously monitors regulatory compliance, security controls, and risk management for organizations.
RegScale 6.32.2.0 expands AI provider support to Azure OpenAI, AWS Bedrock, and OpenAI-compatible endpoints.
Supports Azure OpenAI, AWS Bedrock, and OpenAI-compatible endpoints as AI providers, broadening deployment options for enterprise environments.
CISO Assistant is a one-stop-shop GRC platform for Risk Management, AppSec, Compliance & Audit, TPRM, BIA, Privacy, and Reporting. It supports 150+ 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 v3.19.2 adds asset sync for ServiceNow, reworked risk acceptance workflow, expanded documents management, and Slovak localization.
└──▷ GET THIS VERSION
$ git clone --branch v3.19.2 https://github.com/intuitem/ciso-assistant-community.git
# already have the repo? check out this version:$ git checkout v3.19.2
Supports attaching evidences to data breach records.
Adds extra status options for findings tracking and exceptions.
Adds an autologin page capability.
Adds Slovak (sk) localization.
Adds more customization options for GDPRGDPREU regulation enacted in 2018 that governs how organizations collect, store, and process personal data of EU residents, giving cyber tools legal obligations around data handling and breach notification. Processings.
+8 moreshow less
Supports Import/Export of Representatives via the Data Wizard.
Adds ServiceNow asset sync with schema caching via startup warming and manual refresh.
Adds task template import functionality to the Data Wizard and CLI.
Adds the NCA ECC-2:2024 (Essential Cybersecurity Controls) framework.
CISO Assistant v3.19.1 adds applied control progress as a selectable column and multiple import improvements.
└──▷ GET THIS VERSION
$ git clone --branch v3.19.1 https://github.com/intuitem/ciso-assistant-community.git
# already have the repo? check out this version:$ git checkout v3.19.1
Adds applied control progress as a selectable column in list views, surfacing remediation status at a glance.
Delivers multiple import improvements for loading controls and frameworks.
CISO Assistant v3.19.0 adds SCIMSCIMA standard for creating, updating and removing user accounts and groups across applications. It helps an identity provider keep access in sync when people join, change roles or leave. provisioning, offline AI embeddingembeddingLists of numbers a model produces to represent the meaning of text, images or code, so similar items sit close together. They power semantic search and RAG., managed portals, and admin MFAMFAMulti-factor authentication, requiring proof from different categories, such as a password and a hardware key, to sign in. It reduces reliance on a password alone. reset
└──▷ GET THIS VERSION
$ git clone --branch v3.19.0 https://github.com/intuitem/ciso-assistant-community.git
# already have the repo? check out this version:$ git checkout v3.19.0
Adds SCIM provisioning and IdPIdPA service that authenticates users and supplies identity information to other applications. It is the sign-in authority behind many single sign-on systems. groups mapping for automated user lifecycle management.
Pre-bakes embedding and rerankerrerankerA model that re-scores a first batch of search results against the query and reorders them, more accurately than the fast first-stage search. models into the image for fully offline AI assistant operation (expect larger image size).
Introduces managed portals feature.
Enables admins to reset MFA for other users.
Adds typical evidence and annotation fields to audit/framework Excel exports.
+6 moreshow less
Allows manual implementation groups on dynamic frameworks.
Makes collection creation optional when starting new projects (PMBOK).
Allows admins to disable the partially compliant compliance option.
Helm chart gains CA certificate support, network policy configuration, and AI assistant support.
Completes Italian translation to 100%.
Increases 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. state and nonce length for improved security posture.
CISO Assistant v3.18.3 adds evidence visibility within domains and centralized domain field rendering in forms.
└──▷ GET THIS VERSION
$ git clone --branch v3.18.3 https://github.com/intuitem/ciso-assistant-community.git
# already have the repo? check out this version:$ git checkout v3.18.3
Displays evidence items within their domain view, making it easier to review evidence by organizational scope.
Centralizes folder/domain field rendering across model forms for a consistent UI experience.
Windsurf is an AI-powered code editor that combines an intelligent coding assistant with agentic capabilities to help developers write, edit, and debug code faster.
Devin Desktop v3.5.17 adds inline network policy controls, Fast Context in Local sessions, and instant worktree-backed sessions.
└──▷ USE IT
Restore the status bar in the Agent Command Center after the default changes to hidden.
json
In user settings (JSON), add:
{
"workbench.statusBar.visible": true
}
Adds type-specific icons to @-mention pills in Cascade for clearer context at a glance.
New worktree-backed sessions open instantly and remain interactive while the worktree is created in the background.
Enables configuring a session's network policy and granting or denying network access requests inline from chat, without switching to the web app.
Adds Fast Context support in Devin Local sessions.
The Hooks tab in Devin Local Customizations now lists configured hooks with their source and trigger events.
+5 moreshow less
Adds a command to migrate Windsurf hooks into Devin hooks across every workspace folder.
Adds a subtle timeline navigator for Devin Local sessions.
Devin Local customizations and the sidebar skills count now span all open workspace folders.
View usage now opens eligible Devin users' personal analytics page directly.
The devin.* settings for gitignore access, completion mode, and auto-continue are now honored.
10 details in Everything
└──▷ BREAKING ON UPGRADE
!The status bar is now hidden by default in the Agent Command Center; set "workbench.statusBar.visible": true in user settings to restore it.
Claude Code is an agentic coding tool that lives in your terminal, understands your codebase, and helps you code faster by executing routine tasks, explaining complex code, and handling git workflows - all through natural language commands.
Claude Code v2.1.215 makes /verify and /code-review user-invoked only — no more automatic runs.
└──▷ GET THIS VERSION
$ git clone --branch v2.1.215 https://github.com/anthropics/claude-code.git
# already have the repo? check out this version:$ git checkout v2.1.215
└──▷ TRY IT
Manually trigger a code review after making changes, instead of waiting for Claude to decide to run it automatically.
$ /code-review
Run verification on demand after a refactor to confirm correctness on your own schedule.
$ /verify
The /verify and /code-review slash commands now run only on explicit user invocation, giving practitioners full control over when those checks execute.
1 detail in Everything
└──▷ BREAKING ON UPGRADE
!Claude no longer automatically runs the /verify and /code-review skills; any workflow that relied on Claude triggering these autonomously must now invoke /verify or /code-review explicitly.
Lightweight coding agent that runs in your terminal
Codex CLI gains audio input support, inline visualization links in TUITUIA text-based user interface that runs inside a terminal, using characters and ANSI escape codes to render menus, panes, and widgets without a graphical desktop, letting cyber tools ship rich interactivity with no GUI dependency., and remembered working directory for resumed sessions.
└──▷ GET THIS VERSION
$ git clone --branch rust-v0.145.0-alpha.24 https://github.com/openai/codex.git
# already have the repo? check out this version:$ git checkout rust-v0.145.0-alpha.24
Adds audio input variants (remote and local) to user input protocols, enabling audio to be forwarded to the Responses API.
Renders inline visualization links in the TUI — ::codex-inline-vis{file="..."} directives in assistant Markdown become browser-openable links to generated artifacts.
Enables users to remember the working directory when resuming sessions (tui_resume_cwd).
Allows path-backed agents to be selected from the TUI agent picker.
Renders streamed Markdown incrementally in the TUI for smoother streaming output.
+2 moreshow less
Gates audio history by model input modalities so unsupported models don't receive audio items.
ToolHive is an enterprise-grade platform for running and managing Model Context Protocol (MCP) servers.
ToolHive v0.40.1 adds per-session prompt serving in vMCP and audit logging for authorization denials on the proxy runner path.
└──▷ GET THIS VERSION
$ git clone --branch v0.40.1 https://github.com/stacklok/toolhive.git
# already have the repo? check out this version:$ git checkout v0.40.1
Serves 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. prompts on a per-session basis in VirtualMCP (vMCP), enabling session-scoped prompt isolation.
Audits authorization denials on the proxy runner path, giving operators a log trail of rejected requests.
ToolHive v0.39.0 adds AWS Bedrock compatibility flag to the 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. setup command.
└──▷ GET THIS VERSION
$ git clone --branch v0.39.0 https://github.com/stacklok/toolhive.git
# already have the repo? check out this version:$ git checkout v0.39.0
└──▷ TRY IT
Configure ToolHive to use an AWS Bedrock-compatible LLM endpoint instead of a default provider.
ToolHive v0.38.0 adds Codex App 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. setup support.
└──▷ GET THIS VERSION
$ git clone --branch v0.38.0 https://github.com/stacklok/toolhive.git
# already have the repo? check out this version:$ git checkout v0.38.0
ToolHive v0.37.0 adds podTemplateSpec support for MCPRemoteProxy and exposes --allow-docker-gateway in the workload API.
└──▷ GET THIS VERSION
$ git clone --branch v0.37.0 https://github.com/stacklok/toolhive.git
# already have the repo? check out this version:$ git checkout v0.37.0
└──▷ USE IT
Customize the pod spec for an MCPRemoteProxy deployment — useful when you need node selectors, resource limits, or custom annotations on the proxy pod.
Adds spec.podTemplateSpec support to MCPRemoteProxy, enabling Kubernetes-native pod customization for remote 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. proxies.
Exposes --allow-docker-gateway option in the workload API, allowing workloads to reach the Docker host gateway network.
Adds runtime-stage environment variables to protocol Dockerfiles, enabling environment configuration at container runtime.
Adds MCP conformance CI validation for thv run proxy, improving protocol compliance guarantees.
Validates aud and resource claims in returned ID-JAG JWTsJWTsJSON Web Tokens — an open standard (RFC 7519) for encoding claims as a signed, optionally encrypted JSON object. Cyber tools use them for stateless authentication without server-side session storage., strengthening authentication token verification.
ToolHive v0.35.0 adds rate-limiting observability, HTTP 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. issuer support, and new backend discovery APIs.
└──▷ GET THIS VERSION
$ git clone --branch v0.35.0 https://github.com/stacklok/toolhive.git
# already have the repo? check out this version:$ git checkout v0.35.0
Adds metrics and tracing for rate limiting, enabling observability into throttling behavior.
Supports HTTP upstream OIDC issuers for in-cluster development environments.
Adds ListBackends and LookupBackend methods to the vMCP core interface for programmatic backend discovery.
ToolHive v0.34.0 adds plugin lifecycle management, Claude Desktop gateway support, and Cedar nested-map claim authorization.
└──▷ GET THIS VERSION
$ git clone --branch v0.34.0 https://github.com/stacklok/toolhive.git
# already have the repo? check out this version:$ git checkout v0.34.0
Adds plugin install/list/info/uninstall commands plus a MaterializationAdapter for managing ToolHive plugins.
Adds Claude Desktop as a supported LLMLLMA large language model, a neural network trained on large amounts of text to recognize patterns and generate language. Applications use it for tasks such as answering, summarizing and writing code. gateway client.
Supports nested map claims in Cedar authorization for the act claim, enabling richer policy evaluation.
Allows custom HTTP headers on the OpenAI embeddingembeddingLists of numbers a model produces to represent the meaning of text, images or code, so similar items sit close together. They power semantic search and RAG. client for flexible authentication and routing.
Filters the upstream auth chain via a configurable callback hook (UpstreamFilter) exposed through the authserver.New facade.
+2 moreshow less
Preserves external annotations on operator-managed Kubernetes Services across reconciliation cycles.
Sorts aggregated tools by name for deterministic embedding order.
7 details in Everything
└──▷ BREAKING ON UPGRADE
!The --allow-docker-gateway flag now actively grants host-level network access rather than merely unblocking it, which changes the effective network permissions for containers using this flag.
!The Amp editor extension clients have been removed.
ToolHive v0.33.0 adds rate limiting on VirtualMCPServer and XAA auth strategy CRDCRDA Kubernetes extension that adds a new object type to the cluster, so a tool's configuration is stored, validated and versioned by Kubernetes itself rather than in a file beside it. support.
└──▷ GET THIS VERSION
$ git clone --branch v0.33.0 https://github.com/stacklok/toolhive.git
# already have the repo? check out this version:$ git checkout v0.33.0
Supports configuring rate limits on VirtualMCPServer via the new rate-limit configuration.
Adds XAA authentication strategy support as a CRD 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. server access control.
ToolHive v0.32.0 adds XAA cross-app auth, OBO strategy seam, upstream ID-token propagation, OpenAI-compatible embeddingsembeddingsLists of numbers a model produces to represent the meaning of text, images or code, so similar items sit close together. They power semantic search and RAG., and Origin/DNS-rebind protection.
└──▷ GET THIS VERSION
$ git clone --branch v0.32.0 https://github.com/stacklok/toolhive.git
# already have the repo? check out this version:$ git checkout v0.32.0
└──▷ TRY IT
Allow a browser-based 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. client served from a custom origin to reach ToolHive's 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. transport after the wildcard CORSCORSA browser mechanism that lets a server declare which foreign origins may read its HTTP responses, enforced by the browser to block unauthorized cross-site requests from JavaScript. removal.
$ thv run --transport sse --allowed-origins https://my-web-app.example.com some/mcp-server
Adds XAA (Cross-Application Access) outgoing auth strategy implementing the ID-JAG two-step token exchangetoken exchangeAn OAuth 2.0 protocol extension (RFC 8693) that lets a service trade one token for another with different scopes, audience, or type, enabling delegation and impersonation across service boundaries without re-authenticating the user. (RFC 8693RFC 8693An IETF standard defining the OAuth 2.0 Token Exchange flow, letting a service swap one token for another with different scopes, audience, or type — enabling delegation and impersonation without re-authenticating the user. → RFC 7523RFC 7523An IETF standard that defines how a JSON Web Token (JWT) can be used as a credential to request an OAuth 2.0 access token, letting a cyber tool authenticate to an API without exchanging a username and password at runtime.) for per-backend cross-application access tokensaccess tokensA credential that authorizes a client to use a service or API. It usually has limited permissions and a lifetime; anyone holding it may be able to use those permissions..
Surfaces upstream ID tokens through the auth middleware, consolidating access and ID token retrieval into a single bulk lookup.
Adds the vMCP OBO (on-behalf-of) strategy seam — a new optional OBO field on BackendAuthStrategy with strategy registration and override hook.
Enables the vMCP optimizer to use an OpenAI-compatible embeddingembeddingLists of numbers a model produces to represent the meaning of text, images or code, so similar items sit close together. They power semantic search and RAG. client via optional embeddingProvider/embeddingModel config fields (defaults to TEI, so existing configs are unchanged).
Adds insecureAllowHTTP to EmbeddedAuthServerConfig to allow an http:// issuer for in-cluster VirtualMCPServer deployments, with admission-time validation.
+2 moreshow less
Adds HTTP Origin header validation (MCP 2025-11-25 DNS-rebind protection) and a new --allowed-origins flag on thv run and thv proxy for SSE CORS control.
Operator Helm chart now prints a post-install NOTES.txt with verification commands, a minimal MCPServer example, and documentation links.
7 details in Everything
└──▷ BREAKING ON UPGRADE
!ToolHive now validates the HTTP Origin header and removed Access-Control-Allow-Origin: * from the legacy SSE transport; browser clients on a non-http://localhost origin must pass the new --allowed-origins flag to thv run or thv proxy.
!pkg/vmcp/server.New dropped its discovery.Manager parameter (7 → 6 args) and the pkg/vmcp/discovery package was removed; out-of-tree Go code calling server.New must remove the discovery.Manager argument, ensure Config.Aggregator is non-nil, and pair Config.AuthzMiddleware with Config.Authz or receive ErrInvalidConfig.
Open-Source Evaluation & Testing library for LLM Agents
Giskard Scan adds LIDAR and DeepTeam third-party integrations for broader AI vulnerability coverage.
└──▷ GET THIS VERSION
$ git clone --branch giskard-scan/v1.0.0b3 https://github.com/Giskard-AI/giskard-oss.git
# already have the repo? check out this version:$ git checkout giskard-scan/v1.0.0b3
Integrates LIDAR scan as a third-party scan provider via third_party_scan.
Integrates DeepTeam as a third-party scan provider for additional red-teaming coverage.
2 details in Everything
0 more releases in this issue2 more releases in this issue · 2026-07-13
Giskard Checks v1.0.2b5 adds LiDAR and DeepTeam third-party scan integrations to the vulnerability scanner.
└──▷ GET THIS VERSION
$ git clone --branch giskard-checks/v1.0.2b5 https://github.com/Giskard-AI/giskard-oss.git
# already have the repo? check out this version:$ git checkout giskard-checks/v1.0.2b5
Integrates LiDAR scan into the third-party scan pipeline via third_party_scan.
Adds DeepTeam as a third-party scan integration for agent vulnerability scanning.
Giskard adds crescendo/GCG/do-not-answer/sycophancy scan generators, HarmBench dataset, vulnerability scan API, singleturn/multiturn mode, and new checks.
└──▷ GET THIS VERSION
$ git clone --branch giskard-core/v1.0.1b5 https://github.com/Giskard-AI/giskard-oss.git
# already have the repo? check out this version:$ git checkout giskard-core/v1.0.1b5
└──▷ USE IT
Run a full vulnerability scan against an agent with the new programmatic API, including multi-turn mode.
python
import asyncio
from giskard.scan import vulnerability_scan
async def main():
await vulnerability_scan(
target=my_agent,
description="A customer support chatbot for an e-commerce platform.",
languages=["en"],
)
asyncio.run(main())
Export suite results to Hub-compatible format for upstream reporting or dashboards.
python
result = await suite.run()
hub_payload = result.to_hub_format()
Adds crescendo-like multi-turn adversarial scenario generator for red-teaming 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. agents.
Adds do-not-answer scan generator to probe refusal behavior.
Adds sycophantic scenario generator derived from a knowledge base.
Adds quality scan with knowledge base scenario generator for RAGRAGRetrieval-augmented generation: looking up relevant documents and adding them to the prompt before the model answers. It grounds responses in private or current data without retraining. and agent quality evaluation.
+13 moreshow less
Adds automatic recommendations derived from scan failure categories.
Integrates the HarmBench dataset into the vulnerability scan.
Adds a vulnerability scan API (vulnerability_scan) for programmatic red-teaming.
Adds target_mode flag to switch between single-turn and multi-turn scan execution.
Adds readability NLP metric check to giskard-checks.
Adds match mode to comparison checks for flexible string/value matching.
Adds suite field to SuiteResult for richer result introspection.
Adds to_hub_format() method on SuiteResult to export results to Hub-compatible format.
Adds scenario tag grouping in suite results for organized reporting.
Makes suite failure reporting configurable.
Supports unlimited suite reporting (removes previous reporting cap).
Adds pydantic-settings integration for default model and config management in checks.
Adds fail-fast behavior when instantiating checks that require a missing optional library.
NeMo Guardrails is an open-source toolkit for easily adding programmable guardrails to LLM-based conversational systems.
NeMo Guardrails v0.23.0 adds tool-call rails, HF classifier rails, PIIPIIPersonally Identifiable Information — data that can identify a specific individual, such as names, email addresses, or SSNs. Cyber tools flag or redact it to limit exposure and meet privacy regulations. masking, context bloat detection, and a /v1/checks endpoint.
└──▷ GET THIS VERSION
$ git clone --branch v0.23.0 https://github.com/NVIDIA-NeMo/Guardrails.git
# already have the repo? check out this version:$ git checkout v0.23.0
└──▷ USE IT
Apply Polygraf PII detection and masking to user input before it reaches the 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..
Adds streaming and non-streaming tool callingtool callingA model's ability to answer with a structured request to run a named function or API, which the application executes and returns. It is how agents take actions. to IORails, with local rails that validate model-emitted tool calls and application-returned tool results.
New /v1/checks endpoint on the OpenAI-compatible server runs input or output rails standalone, returning passed, modified, or blocked status without generating a model response.
Adds lightweight Hugging Face classifier rails for input, output, and retrieval, backed by local Transformers, vLLM, KServe, and FMS.
Adds context bloat detection rails to flag oversized, repetitive, low-entropy, or padded input and retrieved content.
Adds Polygraf integration for PII detection and masking across input, output, and retrieval rails.
+4 moreshow less
Replaces Annoy with exact NumPy search as the default embeddingembeddingLists of numbers a model produces to represent the meaning of text, images or code, so similar items sit close together. They power semantic search and RAG. index, removing the native C++ dependency while preserving similarity-threshold semantics.
Extends IORailsOpenTelemetryOpenTelemetryAn 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. support with opt-in content capture and richer request, response, and token-usage span attributes.
LangChain integration adds support for the OpenAI Responses API and Harmony response format models.
Distribution wheel is now approximately ten times smaller by excluding examples and development files.
9 details in Everything
└──▷ BREAKING ON UPGRADE
!Pydantic >=2.5,<3.0 is now required; environments pinned to Pydantic 1.x must upgrade before installing v0.23.0.
A full-stack AI Red Teaming platform securing AI ecosystems via OpenClaw Security Scan, Agent Scan, Skills Scan, MCP scan, AI Infra scan and LLM jailbreak evaluation.
AI-Infra-Guard v4.1.15 adds optional model token in scan APIs, 6 new llama.cpp CVE rules, and 3 new 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. threat detection rules.
└──▷ GET THIS VERSION
$ git clone --branch v4.1.15 https://github.com/Tencent/AI-Infra-Guard.git
# already have the repo? check out this version:$ git checkout v4.1.15
Supports omitting the model token in mcp_scan and ai_infra_scan API calls, falling back to the system default model.
Adds 6 new llama.cpp CVE detection rules to the llama-cpp vulnerability pack.
Adds 3 new MCP threat detection rules covering tool poisoning, credential exfiltration, and command injectioncommand injectionAn attack that makes an application execute operating-system commands supplied through untrusted input. It often happens when input is passed unsafely to a shell..