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.
BloodHound v9.5.0 adds OpenGraphOpenGraphA web metadata protocol, created by Facebook, that lets a webpage declare structured properties—title, image, description, type—so link previews render consistently across social platforms and messaging apps. data quality stats, webhooks UI, markdown-rich entity panels, and new management operation endpoints.
└──▷ GET THIS VERSION
$ git clone --branch v9.5.0 https://github.com/SpecterOps/BloodHound.git
# already have the repo? check out this version:$ git checkout v9.5.0
Adds /data-quality-stats endpoint and aggregations endpoint for querying OpenGraph node and relationship counts by environment.
Adds environment_kind_id and environment_kind_display_name fields to the available-domains API response.
New Webhooks table and webhook details modal on the Delivery Page for managing webhook configurations.
Renders markdown-based KindInfo for OpenGraph extensions directly in the entity panel.
Adds KindInfo to the api/v2/nodes/{node_id} response and to extension upsert operations.
+8 moreshow less
Adds ?include-info query parameter to the v2 relationships API to return KindInfo alongside relationship data.
Introduces start and end management operation API endpoints.
Adds use_raw_object_id feature flag to control raw Object ID handling and case-sensitive search behavior.
Restricts the UPLOAD_ONLY role from accessing certain views, tightening role-based access controlrole-based access controlRole-based access control, where permissions are assigned to roles such as reader or administrator and users receive those roles. It makes access easier to manage than individual permission lists..
Adds read providers permission for finer-grained API access control.
OpenGraph Data Quality page now lists aggregations and environments and renders stats and a node-count table.
Rejects older AzureHound collector versions when the use_raw_object_id flag is enabled.
Improved focus styling across the UI for accessibility.
Veracode is a cloud-based application security platform that helps developers identify and remediate vulnerabilities in their code through static, dynamic, and software composition analysis.
Veracode adds CLI --exclude flag, DASTDASTDynamic application security testing, probing a running application from the outside for vulnerabilities. It checks behavior exposed through the application's interfaces. custom host-to-IP mappings, EU Package Firewall, enhanced Python scanner, and Android 17 / .NET 10 static analysis support.
└──▷ TRY IT
Exclude a legacy JAR or noisy internal package from a Java scan to focus results on application code you own.
Configure a custom host-to-IP mapping for a DAST scan targeting an internal staging environment not resolvable via public DNS.
📍In the Veracode Platform, open your DAST scan configuration, navigate to the Custom Host settings, and add a host-to-IP address mapping for your target host.
Adds --exclude flag to Veracode CLI v2.51.1 to prevent specified Java packages or JAR files from being scanned.
Enables custom host-to-IP address mappings for DAST scans when DNS lookup is unavailable or not desired.
Enhances Python static scanner with taint tracking, cross-file analysis, and pattern recognition; adds framework support for PySpark, Flask, Django, and FastAPI.
Adds support for Android 17 in the Static Analysis Engine v2026.07.
Enhances support for .NET 10 in the Static Analysis Engine v2026.07.
+4 moreshow less
Expands third-party library detection to include Fluxnova.
Expands support for pnpm modules to improve dependency resolution in static analysis.
Improves JSP and JSF Expression Language modeling with more complete POJO data, potentially surfacing previously undetected findings.
Enhances SQL dialect packaging support in Veracode CLI v2.51.1.
Snyk finds and fixes security vulnerabilities in code, open-source dependencies, and containers throughout the development pipeline.
Snyk DASTDASTDynamic application security testing, probing a running application from the outside for vulnerabilities. It checks behavior exposed through the application's interfaces. gains AI-powered authenticated scan setup and native Bruno collection support for API targets.
└──▷ HOW TO FIND IT
Set up an authenticated DAST scan without manually recording login flows — use the AI-powered configuration to handle login sequences and 2FA in minutes.
📍In the Snyk console, go to your DAST target settings, select Automatic Target Authentication Configuration, enter your credentials in the dedicated credential fields, and optionally provide natural-language instructions for complex login scenarios.
Onboard a Bruno-based API collection as a Snyk API Target without converting files to Postman or OpenAPIOpenAPIA vendor-neutral specification for describing REST APIs in a machine-readable format, maintained by the OpenAPI Initiative. Cyber tools use it to auto-generate clients, documentation, and validation from a single source of truth..
📍In the Snyk console, navigate to API & Web › API Targets › Create Target, select Bruno Collection, and upload your collection as a ZIP file or folder.
Adds Automatic Target Authentication Configuration for DAST: AI-driven setup automatically generates login sequences, handles 2FA, and maps CSS selectors, with natural-language instructions for complex scenarios.
Supports native Bruno collection imports for API Targets — upload via ZIP or folder without converting to Postman or OpenAPI formats first.
Syncs Bruno collection environments automatically and warns when variables are missing required secret values during import.
Portkey is an AI gateway and observability platform that helps developers manage, monitor, and optimize LLM API calls with routing, caching, and reliability features.
Portkey March 2026: vault-backed credentials, weekly rate limits, cross-cloud GCS logging, and four new guardrail integrations.
└──▷ USE IT
Apply Zscaler AI Guard detection policies to all 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 by wiring in your existing policyId.
Adds Secret References to pull credentials from AWS Secrets Manager, Azure Key Vault, or HashiCorp Vault at runtime via secret_mappings, keeping secrets out of Portkey config.
Supports weekly budget and usage windows (rpw) for rate limits, enabling caps aligned to weekly planning and review cycles.
Adds endpoint-scoped rate limits so chat completions, 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 admin routes can carry independent limits instead of a single global rule.
Enables GCP Workload Identity Federation log delivery from AWS-hosted gateways to GCS via GCP_WIF_AUDIENCE and GCP_WIF_SERVICE_ACCOUNT_EMAIL, eliminating long-lived cross-cloud keys.
Allows organization admins to include archived workspaces in analytics graphs and summaries; saved filters now accept workspace slugs alongside IDs.
+11 moreshow less
Integrates Zscaler AI Guard as a guardrail partner, applying Zscaler Detection Policies to LLM inputs and outputs via beforeRequestHook and afterRequestHook with a required policyId.
Integrates Akto Agentic Security as a guardrail partner for scanning LLM inputs and outputs for 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. and sensitive data leakage.
Adds customHost to the Bedrock Guardrails plugin, routing guardrail evaluation to private or regional Bedrock-compatible endpoints.
Adds required metadata key–value guardrails that block requests missing specified metadata fields before they reach the model.
Adds DeepInfra provider support with 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. (tools, tool_choice, parallel_tool_calls), completions, and embeddings endpoints.
Expands DeepSeek support: deepseek-chat gains tools, tool_choice, and stream_options; deepseek-reasoner maps reasoning_effort to thinking mode and returns reasoning_content in streams.
Extends Bedrock batch inference to support embeddings jobs alongside existing chat completions batch patterns.
Maps Portkey metadata to Vertex AI resource labels and adds enterprise search grounding via enterpriseWebSearch / enterprise_web_search with separate cost attribution.
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.
Adds Azure AI Foundry rerank support for Cohere rerank models (e.g., cohere.Cohere-rerank-v4.0-pro), with the gateway stripping the cohere. prefix for the provider.
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 for inference and embeddings, with 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 custom resource attributes.
Wiz Code is a developer security tool that identifies and remediates code vulnerabilities and misconfigurations in applications.
Wiz introduces Atlas, an autonomous AI SASTSASTStatic application security testing, analyzing source code or compiled code for vulnerabilities without running the application. It helps developers find problems before deployment. system that validates every finding with a working exploit and scans pipelines for Secure by Design compliance.
Introduces Atlas, an autonomous AI vulnerability research system that uses specialized agents to map 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., hunt in parallel, and adversarially validate findings — each confirmed with a dynamically generated, working exploit.
Adds AI SAST at scale: Atlas performs continuous, cost-efficient deep code scanning across entire enterprise codebases using a code property graph (CPG) for call- and data-flow–grounded analysis.
Adds shift-left tracing that links vulnerabilities discovered in production back to their originating code repository or container image, enabling fixes at the source.
Extends pipeline scanning to AI application configurations and model deployments, in addition to container images, software packages, 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., to enforce Secure by Design guardrails.
StepSecurity is a platform that secures CI/CD pipelines by detecting and preventing supply chain attacks, hardening GitHub Actions workflows, and monitoring runtime behavior.
Enforce an internal ban on all 4.x versions of lodash and an entire scoped namespace across your npm policy.
📍In the Secure Registry console, go to Policy › npm › Custom Block List and add entries: lodash@4* and @rejected-scope/*, then save.
Route all Maven artifact downloads through Secure Registry for evaluation instead of Maven Central.
📍In the Secure Registry console, go to Setup Guide › Registry selector › Maven, select your integration path (e.g., Direct), then point your Maven client at https://registry.stepsecurity.io/java as the mirror URL.
Ensure npm package tarballs download through Secure Registry rather than the public npm registry.
📍In the Secure Registry console, go to npm Registry Settings and enable the Tarball URL Rewriting toggle.
Adds Maven (Java) ecosystem support (Beta) to Secure Registry — route all Maven metadata and artifact downloads through https://registry.stepsecurity.io/java with Cooldown Period, Compromised Packages, and Custom Block List controls at launch.
Supports four Maven integration paths: JFrog Artifactory, Google Artifact Registry, Sonatype Nexus Repository, and Direct, with setup steps switchable via a new Package manager selector on the Setup Guide tab.
Introduces Custom Block List, a new per-ecosystem control that blocks specific packages or versions using glob patterns (e.g., lodash@4*, @scope/*), evaluated independently of Cooldown, Compromised Packages, and TyposquattingTyposquattingUsing a name close to a trusted domain or software package to catch spelling mistakes. Attackers use it to lure people or installers to a malicious lookalike. Protection.
Adds Tarball URL Rewriting setting for npm, rewriting tarball URLs in metadata responses so package downloads route through Secure Registry instead of the public npm registry.
DataDome Python ASGIASGIA Python async web server interface specification that defines how servers and applications exchange HTTP and WebSocket messages; cyber tools use it to serve high-concurrency APIs without blocking on I/O. module 1.2.0 adds configurable 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. endpoint support and expands cookie length to 512 chars.
Adds mcp_endpoint attribute to DataDomeMiddleware to specify a custom MCP endpoint path (defaults to /mcp).
Increases DataDome cookie length limit from 128 to 512 characters to support future capabilities.
GitGuardian SaaS is a secrets detection platform that scans code repositories and developer workflows to identify and remediate exposed credentials and sensitive data.
GitGuardian adds a dedicated GitHub Check Runs dashboard page with filtering, KPIs, re-run controls, and incident linking.
└──▷ HOW TO FIND IT
Audit scan coverage and unblock a stuck pull request by finding and re-running a failed check run without leaving GitGuardian.
📍In the dashboard, go to Perimeter › GitHub check runs. Filter by repository and set Status to Failed, locate the run for the relevant commit SHA, then click Re-run.
Jump straight to remediation when a check run catches a secret, without switching to GitHub.
📍In the dashboard, go to Perimeter › GitHub check runs. Find the check run with detected secrets, then click the linked incident to open the incident details and begin remediation.
New GitHub Check Runs page in the Perimeter section surfaces every check run with status, repository, commit SHA, pull request, and timing in one place.
Filter and search check runs by repository, status, or date — including workspaces with hundreds of thousands of runs.
KPI strip at the top of the Check Runs page summarizes activity to surface failures and trends at a glance.
Re-run or skip a check run directly from the dashboard without switching to GitHub.
Navigate from a check run directly to linked incidents when secrets are detected, enabling faster remediation.
+2 moreshow less
Fine-grained GitHub personal access tokenspersonal access tokensA 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. can now be used to monitor private members of a GitHub organization in the Public Perimeter (previously limited to classic tokens).
Sources health management extended to Bitbucket Cloud and Bitbucket Data Center: GitGuardian now auto-pauses ingestion on unreachable sources, auto-resumes on recovery, and surfaces actionable recovery steps.
Infisical is the open-source platform for secrets, certificates, and privileged access management.
Infisical v0.162.14 adds Google Workspace service template, Windows domain account rotation, enriched secret reminder emails, and instance version update checks.
└──▷ GET THIS VERSION
$ git clone --branch v0.162.14 https://github.com/Infisical/infisical.git
# already have the repo? check out this version:$ git checkout v0.162.14
Adds Google Workspace as a proxied service template (available as a popular template in the service proxy catalog).
Tracks and displays a last-used timestamp for proxied services.
Adds an instance version update check with in-UI display so self-hosted admins can see when a newer release is available.
New machine identity creation sheet includes policy templates to accelerate identity provisioning.
Enriches secret reminder emails with secret details and a direct View secret link.
+1 moreshow less
Adds PAMPAMBrokering administrator sessions through a gateway instead of handing out standing credentials. Access is granted for one session, to one target, and recorded. domain account dependency detection and Windows credential rotation support.
Edit and replay captured network traffic at arbitrary speeds — a suite of pcap tools for *NIX and Windows.
tcpreplay 4.6.0 adds AF_XDPAF_XDPA Linux kernel interface that hands raw packets to a userspace program straight off the network card, skipping the kernel's networking stack. It is how userspace tools reach line rate on 100G links. and io_uring injection backends, --loss simulation, --raw socket path, and a static libtcpreplay library.
└──▷ GET THIS VERSION
$ git clone --branch v4.6.0 https://github.com/appneta/tcpreplay.git
# already have the repo? check out this version:$ git checkout v4.6.0
└──▷ TRY IT
Saturate a 1GigE link using the AF_XDP kernel-bypass backend; fall back automatically if the adapter lacks native XDPXDPA Linux kernel interface that hands raw packets to a userspace program straight off the network card, skipping the kernel's networking stack. It is how userspace tools reach line rate on 100G links. support.
$ tcpreplay --xdp-i eth0 capture.pcap
Benchmark AF_XDP on a specific queue without silent fallback masking results — fail loudly if XDP can't be used.
Simulate a lossy link by randomly dropping 10% of packets during replay to stress-test application resilience.
$ tcpreplay --loss 10 -i eth0 capture.pcap
Adds --xdpAF_XDP injection backend, now reaching ~941 Mbps (305k pps) line rate on 1GigE hardware — a 23× improvement over the previous implementation.
Adds --xdp-queue to bind the AF_XDP socket to a specific adapter queue instead of always using queue 0.
Adds --xdp-no-fallback to fail hard rather than silently fall back to the default injector — useful for benchmarking where a silent injector change would invalidate results.
Adds --xdp-batch-size for tuning AF_XDP batch depth on very high-speed links (100GigE+) with small packets.
Adds --io-uring injection backend that batches sends 64 deep via Linux io_uring, reducing per-packet syscall count 64× for ~11% less CPUCPUThe main general-purpose processor in a computer. It runs application instructions and coordinates work with memory and other hardware. per packet.
+5 moreshow less
Adds --rawPF_INET/SOCK_RAW injection path that routes packets through the kernel IP stack (routing, netfilter); IPv4-only, rebuilds L2 framing.
Adds --loss flag for random percent packet-loss simulation (0–100%), enabling testing against non-ideal network conditions.
Adds support for replaying onto raw IP (L3-only) interfaces such as WireGuardWireGuardA modern VPN protocol and Linux kernel module, maintained as an open-source project, that establishes encrypted tunnels using minimal code and fast cryptography, giving cyber tools a lightweight, high-performance transport layer. and tun, which have no L2 header.
Installs the replay engine as libtcpreplay, a standalone static C library.
Adds CMake build system (cmake -B build && cmake --build build) as an alternative to autotools, with full feature parity.
Rapid7 InsightCloudSec is a cloud-native security platform that continuously monitors, detects, and automatically remediates security risks and compliance violations across multi-cloud environments.
InsightCloudSec 26.7.28 adds AI Root Cause Analysis for runtime findings, HITRUSTHITRUSTA US-based certification framework for healthcare data security, maintained by the HITRUST Alliance, that maps controls from HIPAA, NIST, and ISO into a single assessable standard tools can target to prove compliance.CSFCSFThe NIST Cybersecurity Framework, a set of outcomes for managing cybersecurity risk. Organizations use it to describe their current posture, priorities and improvement goals. v11.7.1 compliance packs, and AWS Tag Policy resource harvesting.
└──▷ HOW TO FIND IT
Investigate why a container runtime detection fired by reviewing AI-generated root causes and when the underlying posture risks first appeared.
📍In InsightCloudSec, open a container runtime detection finding › AI Analysis panel › Root Cause Analysis tab to view the AI summary, Prevention Window, Primary Causes with confidence scores, and Contributing Factors.
Adds a Root Cause Analysis tab to the AI Analysis panel for container runtime detection findings, surfacing an AI-generated summary, a Prevention Window showing when posture risks were first detected relative to the runtime incident, Primary Causes with confidence scores, and Contributing Factors.
New HITRUST CSF v11.7.1 e1 Essential and i1 Intermediate compliance packs available.
Adds harvesting support for AWS Organizations Tag Policy resources across AWS, AWS GovCloud, and AWS China environments via the new TagPolicyHarvester, enabling visibility into tag enforcement policies.
Qualys TotalCloud is a cloud security platform providing vulnerability management, compliance monitoring, and threat detection across cloud infrastructure.
Qualys TotalCloud adds automated vulnerability scanning for AWS Lambda, extending serverless security beyond CSPMCSPMCloud security posture management, checking cloud resources for risky settings and policy violations. It helps teams find exposed services, overly broad permissions and other configuration gaps. to workload-level risk.
└──▷ HOW TO FIND IT
Deploy the Qualys Lambda scanning integration to your AWS environment to enable automated, event-driven vulnerability scanning across accounts and regions.
📍Deploy the provided CloudFormation template in your AWS account (same-account or cross-account) to configure event-driven Lambda function scanning. Once deployed, Qualys automatically discovers and scans Lambda functions on creation or update — no manual triggering needed.
Adds automated vulnerability scanning for AWS Lambda functions, detecting vulnerable open-source packages and outdated libraries inside function code — beyond what CSPM posture checks can see.
Enables event-driven scanning that triggers automatically when Lambda functions are created or updated, with no manual intervention required.
Supports cross-account and multi-region AWS Lambda environments, including same-account and cross-account configurations deployed via CloudFormationCloudFormationAn AWS service that lets you describe and provision cloud infrastructure as declarative YAML or JSON templates, so a cyber tool can reproducibly spin up or tear down its entire AWS environment with a single API call. templates.
Combines Lambda vulnerability findings with cloud security context (permissions, exposure) to power hyper-prioritization via automated TruRisk Scoring and unified attack pathattack pathA possible sequence of steps an attacker could take to reach a valuable target. It connects weaknesses, permissions and reachable systems rather than treating each finding separately. analysis merging CSPM and real-time detections.
Aikido Security is a developer-first security platform that identifies and fixes vulnerabilities in code, dependencies, and infrastructure.
Aikido adds DSPMDSPMData security posture management, locating sensitive data and checking how it is protected and who can reach it. It helps teams prioritize risks around their data stores., backport library patches, code quality scans, and new agent/IDE releases in one drop.
└──▷ HOW TO FIND IT
Point DSPM at a repository to discover where sensitive data is exposed in code — no database credentials or access grants needed.
📍In the Aikido console, navigate to your repository, then open the Data Exposure Audit (DSPM) scan to review sensitive data flow findings and remediate at the source.
Merge a backported CVE fix for an existing dependency version without upgrading to a new major version.
📍In the Aikido console, go to Aikido Libraries, select an open-source CVE affecting your current dependency version, and open the generated PR to apply the backported patch.
New Data Exposure Audit (DSPM) scans repos to track how sensitive data flows and where it gets exposed — no database connections or data access grants required.
New Aikido Libraries capability delivers backported CVE fixes for the exact dependency version already in use, without requiring major-version upgrades or refactors.
New Code Quality Repository Scans surface existing code quality issues across repos to help identify and prioritize technical debt over time.
Noise reduction for container images in Cloud scans.
Bitbucket Server support added for PR Checks.
+7 moreshow less
Kubernetes registry option added for public image scanning.
PR/MR gating by severity now available on GitHub for Code Quality findings.
Device Protection adds support for JetBrains Zen.
Firewall agents now available for PHP, Ruby, Golang, and Node.
IDE extensions updated with new releases for VSCode, JetBrains, and Visual Studio.
Health score added for the Maven ecosystem in Aikido Intel.
Bulk assignment of cloud accounts to teams available in Settings.
XFA is a developer security tool that scans and monitors open-source dependencies for vulnerabilities, helping teams manage software supply chain risks efficiently.
XFA desktop app gains a full security dashboard so end users can see every check and self-remediate without IT.
New full desktop app UI (macOS, Windows, Linux): users can open the menu-bar icon to see a complete device security overview.
Per-check pass/fail detail view shows every security check XFA runs with step-by-step remediation guidance for end users.
Huntress is a managed threat detection and response platform that helps MSPs and enterprises detect and respond to security threats.
Huntress Managed ITDRITDRIdentity Threat Detection and Response: a discipline focused on detecting and responding to attacks that abuse identity systems such as Active Directory or IAM platforms, giving cyber tools visibility into credential-based intrusions. now infers Microsoft 365 home country from tenant data, eliminating false-positive Unexpected Country escalations.
Automatically infers a Microsoft 365 identity's home country from the tenant when no Entra usage location is set, requiring zero partner configuration.
Inferred country value serves as the Expected Country for Unexpected Country escalation evaluation, reducing false-positive alerts for users logging in from their home country.
Managed ITDRITDRIdentity Threat Detection and Response: a discipline focused on detecting and responding to attacks that abuse identity systems such as Active Directory or IAM platforms, giving cyber tools visibility into credential-based intrusions. now throttles and prioritizes escalations during a new tenant's first week to reduce onboarding noise.
Adds onboarding escalation throttle to Managed ITDR, pacing and prioritizing alert flow during a new tenant's first week while behavioral baselines build.
Microsoft Defender is an antivirus and security tool protecting Windows devices from malware, viruses, and cyber threats.
Microsoft Defender's Project Perception launches public preview with coordinated red/blue/green AI agents for closed-loop autonomous defense.
Introduces Project Perception, a continuously learning defense system that coordinates three classes of specialized AI agents — red team (attack pathattack pathA possible sequence of steps an attacker could take to reach a valuable target. It connects weaknesses, permissions and reachable systems rather than treating each finding separately. discovery), blue team (investigation and risk reasoning), and green team (corrective action and hardening) — operating in a closed-loop across identities, endpoints, applications, data, clouds, and AI systems.
Launches software vulnerability management as the first Project Perception scenario, integrating MAI-Cyber-1-Flash into MDASH (software vulnerability multi-model agent team) for automated vulnerability discovery and prioritization.
MDASH with MAI-Cyber-1-Flash achieves 96% on the CyberGym benchmark (+12 points above the previous Mythos configuration) while delivering ~50% cost reduction versus the current production MDASH configuration.
Microsoft Security provides comprehensive cybersecurity solutions, threat protection, and security monitoring tools for organizations and enterprises.
Microsoft launches Project Perception agentic security system and the External Red Team Alliance (EXTRA) for global AI safety research.
Introduces Project Perception, a new agentic security system with a multi-layer Cyber Stack — signals/sensors, security context, multi-model architecture, agent harnessagent harnessThe software around an AI agent that manages its tools, execution loop, state and limits. It determines how the model's proposed actions are run and checked., and actuators — designed to continuously perceive, reason, and act at machine speed.
Adds a continuously updated security context layer that transforms signals, threat intelligence, and asset/identity/relationship data into a shared, near-real-time representation for AI agents, reducing token cost and improving reasoning accuracy.
Adopts a multi-model architecture that dynamically selects the best-fit model (frontier or specialized) per security task, optimizing for quality, latency, reliability, and cost without locking customers to a single model.
Introduces actuators that translate AI reasoning directly into real-time protections and actions across integrated security workflows, closing the gap between insight and enforcement.
Launches EXTRA (External Red Team Alliance), a formalized global program funding 18 university labs across six continents with unrestricted gifts to advance independent AI safety and red-teaming research.
Lookyloo is a web interface that allows users to capture a website page and then display a tree of domains that call each other.
Lookyloo v1.40.1 adds WebSocketWebSocketA protocol that upgrades an HTTP connection into a persistent, full-duplex channel, letting a server push data to a client without repeated polling. Cyber tools use it for real-time event streaming and live command output. stream visibility, private captures with token access, and background tree-building scripts.
└──▷ GET THIS VERSION
$ git clone --branch v1.40.1 https://github.com/Lookyloo/lookyloo.git
# already have the repo? check out this version:$ git checkout v1.40.1
└──▷ HOW TO FIND IT
Restrict a sensitive capture so only admins or a shared temporary token can view it.
📍In the capture view, open the visibility form and set the capture to Private; share the generated temporary token with authorised recipients.
Adds private captures mode — captures visible only to admins or holders of a temporary token.
Adds a UI form to change the visibility of an existing capture (e.g. public → private).
Adds background scripts to build trees and index archived captures without blocking the main process.
Intruder is a cloud-based vulnerability scanner that continuously monitors and identifies security weaknesses in internet-exposed systems, APIs, and web applications.
Intruder launches AI Pentesting: on-demand web app pentests with PDF reports, whitebox code repo analysis, and logic-flaw detection.
└──▷ HOW TO FIND IT
Start an AI-driven pentest against your web app directly from the Intruder portal.
📍In the portal, go to AI Pentesting in the left nav, click Start new pentest, then choose the AI Pentest option.
New AI Pentesting capability delivers full web app penetration tests on demand, producing a PDF report with executive summary, severity ratings, reproduction steps, and remediation guidance.
Detects access control, business logic, and auth bypass vulnerabilities that require multi-step chaining — issues missed by static or purely dynamic scanning.
Allows practitioners to define custom entry points, credentials, and application context to scope tests to their real environment.
Censys is an internet intelligence platform that continuously scans and indexes internet-connected devices, helping security teams discover and monitor exposed assets and vulnerabilities.
Censys adds OAuth 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 setup, 10 new software fingerprints, and 23 new protocol/application scanners.
└──▷ HOW TO FIND IT
Connect the Adversary Investigation MCP server to your account using the new OAuth workflow instead of manual credential setup.
📍In the Platform web console, go to Connected Applications, then follow the OAuth workflow to authorize the Adversary Investigation MCP server.
New Connected Applications page in the Platform web console lets users review and manage OAuth connections.
Adds 10 new software fingerprints covering Apple HTTPServer, Control Web Panel, Dynatrace, LinuxServer SWAG, Meteor, Microsoft ASP.NET MVC, Microsoft WCF, Oracle Mojarra, Weaver e-cology, and ZKoss ZK Framework.
Adds 23 new protocol, application, and endpoint scanners including CORTEX_XDR_P2P, DENON_HEOS, DOTNET_NEGOTIATE_STREAM, ELASTICSEARCH_TRANSPORT, GIOP, INFLUXDB, ISC_DHCP_OMAPI, JUPYTER, KCODES_NETUSB, KEYCLOAK, LANGFLOW, MYSQLX, NGINX_PROXY_MANAGER, PERFORCE_P4D, PROXMOX_VE, QDRANT_GRPC, R1SOFT_BUAGENT, SYNCTHING_BEP, SYNOLOGY_DSM, TARANTOOL, TELEXPER_TLXP, TWAMP_CONTROL, and VAULT.
Ory is an open-source identity and access management platform providing authentication, authorization, and user management infrastructure for developers building secure applications.
Passbolt is an open-source password manager designed for team collaboration, enabling secure password sharing and management across organizations.
Passbolt 5.14 adds smarter credential suggestions, stronger 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. policy enforcement, and email-free SSOSSOSingle sign-on, using one identity provider to sign in to multiple applications. It reduces separate passwords and centralizes account and access management. setup and recovery.
Improves visibility over shared access for resources.
Delivers more relevant in-context credential suggestions.
Strengthens MFA policy enforcement.
Expands email-free SSO setup and recovery workflows.
ConductorOne is an identity governance platform that automates access management, provisioning, and compliance across enterprise applications and infrastructure.
ConductorOne adds AI agent policy steps, group/role access reviews, per-user 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. auth, and scoped personal API keys.
└──▷ HOW TO FIND IT
Delegate a policy step to the C1 AI agent so it automatically evaluates access requests and logs each decision for audit — no human reviewer needed for routine approvals.
📍In the console, go to Policies › [your policy] › Edit, add or select a workflow step, choose Assign to agent, enter any custom instructions, and save. Audit log entries for that step will appear under the C1 System agent identity with the AI-agent action label.
Scope a personal API key to read-only to limit blast radius when integrating with a third-party reporting tool.
📍In the console, go to Settings › Personal API Keys › Create API Key, open the Limit Scopes selector, choose the read-only role, and generate the key.
Run a quarterly access review that covers group and role grants, not just individual user grants, to satisfy broader compliance requirements.
📍In the console, go to Access Reviews › Create Campaign, open the Scope tab, find the Grant types selector, choose Users and resources, set a custom recurrence interval (e.g., every three months), configure the campaign duration, and launch.
Assign a workflow step in a request or review policy to the C1 AI agent, which evaluates each request using built-in guidance plus custom instructions; every evaluation is attributed to the C1 System agent identity, labeled as an AI-agent action in the audit log, and recorded as a reviewable conversation.
Request catalog search now includes an Entitlements matching section showing entitlements that indirectly grant access to a searched resource (e.g., a GitHub team whose repo permissions include the target repo), with a View access graph link showing the full grant chain.
Access review campaigns can now review grants held by groups or roles in addition to users, using a new Grant types selector (Users only, Resources only, or Users and resources) on the Scope tab.
Campaign templates now support custom recurrence intervals (e.g., every two days, every three months) and configurable campaign durations, beyond the previous fixed presets.
IAMIAMIdentity and Access Management — a framework of policies and services for controlling who can authenticate to systems and what resources they are authorized to use, letting cyber tools enforce least-privilege and audit access. policy documents on AWS resources now render as a readable, collapsible JSON tree via a View policy button on resource-level access reviews and in task details.
+10 moreshow less
C1 groups can now be added to a request catalog or access profile, making group membership requestable through the standard approval flow.
Requests for entitlements created ahead of a connector sync (e.g., via Terraform) now queue and wait rather than fail, display a status reason, resume automatically after the next sync, and fall back to manual provisioning after 24 hours.
A new Retry failed provisioning action on access tasks lets admins retry connector provisioning in place without canceling the task and losing collected approvals.
Step-up (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.) approvals that also require a comment now prompt for identity-provider verification in a single flow and complete once verified.
C1 AI assistant reports now include a View lineage button showing the exact queries behind each number, verified against the execution log, and copyable for audit workpapers.
C1 AI now renders multiple requested charts as an aligned grid with individual export and source links, and can respond with composed reports combining stat cards, a trend chart, and a data table.
A new Generate action in the CELCELCommon Expression Language, an open-source expression language maintained by Google, used to embed lightweight, safe policy evaluation logic into tools without running a full scripting engine. editor converts a plain-English description directly into a CEL condition expression in one step.
Admins can now require per-user authentication (OAuth2OAuth2An 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., bearer tokenbearer tokenA credential that grants access to whoever presents it, without a separate proof of possession. Protecting the token matters because a stolen copy can be used by someone else., custom header, or basic auth) for external MCP servers instead of sharing a single admin credential.
Tool enablement settings (require enablement or auto-enable) can now be configured per MCP server, independently of the tenant-wide AI governance default.
Any user can now restrict a personal API key to a narrower role (e.g., read-only) using the Limit Scopes selector at key creation, without requiring Super Admin access.
SailPoint is an identity security platform that helps organizations manage and govern user access to applications, data, and systems across the enterprise.
SailPoint introduces an AI-guided connectivity agent wizard to configure Web Services sources without deep API expertise.
Adds a connectivity agent: an AI-powered, wizard-based setup experience for configuring Web Services connector sources directly within the SailPoint application onboarding workflow.
Supports AI-recommended configuration for authentication endpoints, account aggregation with data preview, pagination rules for large datasets, attribute mapping for accounts and entitlements, and identity-to-account correlation.
Enables detection of compatible REST and 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. public APIs for any target application, compressing multi-day Web Services configuration into hours.
Includes provisioning endpoint recommendations covering Get Object, Create Account, Account Update, and Add/Remove Entitlement.
Each AI recommendation is validated in-UI with a live data preview before the administrator accepts — no changes are applied without explicit approval.
+1 moreshow less
Standard manual Web Services configuration flow remains available alongside the new wizard for teams requiring full control.
Clerk is a user authentication and identity management platform that provides developers with pre-built UI components, APIs, and SDKs for sign-up, sign-in, and user management.
Clerk introduces composable, experimental building-block APIs for UserProfile and OrganizationProfile pages.
Progress ShareFile is a cloud-based file sharing and storage platform enabling businesses to securely send, receive, and manage documents and client communications.
ShareFile adds Tasks workspace, Request List export, Outlook signature requests, session management, AI document validation, and a Security Insights dashboard.
New Export a Request List capability lets users export Request Lists to Excel (.xlsx.xlsxThe file format used by Microsoft Excel for spreadsheets, based on the Open XML standard. Cyber tools use it to ingest or export structured tabular data in a format analysts already work with.) or CSV (.csv) from the three-dot menu, reflecting currently active columns.
New ShareFile Tasks provides a centralized workspace for viewing and managing work items across projects in a single location.
New Create Signature Request in ShareFile for Outlook Online add-in lets users send documents for e-signature or self-sign directly from Outlook without leaving the inbox.
ShareFile for Windows now provides access to Projects, Client Hub & Solutions files and folders for all Premium plan users (v26.03.12.0+).
New Bulk client entity creation in Client Hub allows admins and employees to create multiple client entities and primary contacts at once.
+16 moreshow less
New Transparent Automation Progress surfaces real-time status, progress, and outcomes of long-running process automations directly in the ShareFile UI.
New Sessions page under Personal Settings lets users view all active sessions; admins can manage any user's sessions from their profile; users can sign out of all sessions on password change.
New Security Insights dashboard provides admins a visual security dashboard with real-time alerts, event breakdowns, geographic context, remediation statuses, and user-level insights.
New ShareFile authentication one-time codes give admins a new secure authentication method for client users.
New Social Sign-In Microsoft integration allows admins to enable Microsoft as a social sign-in provider for client contacts.
New Client Hub Custom Fields enable admins to add unique data fields tailored to their operational needs.
Outlook Online Email Recipients Sign-In now requires recipients to authenticate before accessing shared content, with account activation for new users.
New ShareFile for Clients mobile applications (iOS and Android) for secure file exchange with businesses and professionals.
New Closed Project Behaviour lets project owners close projects, making all content read-only and marking all cases as closed for data retention compliance.
New Create Payment Requests within ShareFile Engagements allows service providers to send payment requests directly to clients from ShareFile.
New In-App Security Notifications deliver security alerts inside the ShareFile UI in addition to email.
ShareFile for Windows (25.7.4.0) now supports ARMARMA family of RISC processor architectures designed by Arm Holdings and licensed to chip makers worldwide; tools target it to run natively on mobile, embedded, and increasingly server and desktop hardware.-based processor devices.
New AI Driven Document Validation automatically verifies that client-uploaded files match document request specifications.
Redesigned Document Request List experience with a new layout for faster reviewing, uploading, and managing of documents.
Storage Zone Controller 6.0 is rebuilt on modern .NET Core with significant performance, security, and usability improvements and streamlined configuration and deployment.
Client Hub (Beta) allows administrators to manage client lists efficiently within ShareFile.
21 details in Everything
└──▷ BREAKING ON UPGRADE
!Upgrading to ShareFile for Mac 26.04.20 requires all users to re-authenticate.
!Upgrading to ShareFile for Windows 26.3.12.0 requires all users to re-authenticate.
!ShareFile for Windows 25.6.5.0 is now signed with Progress Software Corporation's digital certificate; environments restricting software by publisher may block this update until Progress Software Corporation is added to the trusted publishers list.
DNSFilter is a cloud-based DNS security platform that blocks malicious websites, filters unwanted content, and protects networks from cyber threats in real time.
DNSFilter adds Brave SafeSearch enforcement, AppAware Sites filter, clickable MSPMSPA managed service provider, a company that operates IT services for customers. It may manage systems, networks, backups or other ongoing technology needs. widgets, and Query Log Uncategorized filtering.
└──▷ HOW TO FIND IT
Enforce SafeSearch on Brave (and all other supported engines at once) in a Filtering Policy.
📍In the console, go to Filtering Policies › select a policy › Privacy tab › SafeSearch section › click Enforce All to enable every supported engine including Brave.
Find unclassified DNS queries that may need custom categorization or blocking.
📍In the console, go to Query Log › add a filter for Uncategorized to display only queries DNSFilter has not yet classified.
Adds Sites filter to AppAware Reporting, giving Core plan and mixed Core/Pro environments query data scoped by site.
Supports SafeSearch enforcement on Brave browser in Filtering Policies, joining Google, Bing, DuckDuckGo, Ecosia, and Yandex.
Adds Enforce All shortcut in the redesigned SafeSearch policy editor to enable every SafeSearch engine at once; enforcement toggles are no longer nested under Block all search engines.
Enables filtering DNS Query Log by Uncategorized to surface unclassified queries.
Makes Collections and Users widgets on the MSP Overview page clickable, linking directly to their respective management pages.
Immuta is a data security platform that automates access control, privacy policy enforcement, and compliance monitoring across cloud data environments.
Immuta adds separate access/masking request forms, Redshift viewless auto-discovery, managed groups for policies, 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
Write a subscription policy that allows every registered identity (agents and users) to subscribe to non-sensitive data without manually maintaining group membership.
📍In the Immuta policy editor, create a subscription policy and set the target group to the built-in managed group All Identities to automatically include all agent and non-agent user accounts.
Adds separate access and masking exception request forms for data products, allowing stewards to configure independent questions and data use agreements per request type.
Expands automated sensitive data discovery to Immuta's Amazon Redshift viewless integration, automatically tagging matching columns using regex, dictionary, and column-name regex identifiers.
Introduces three built-in Immuta managed groups — All Agents, All Users, and All Identities — available for authoring subscription and data policies without manual group management.
Adds 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 authentication support for Immuta's Alation integration alongside existing API key authentication, eliminating recurring token refresh requirements.
Extends Snowflake manual policy resync to cover both data policies and subscription policies when triggered from a data source health check.
+2 moreshow less
Enables Unity Catalog tag ingestion for 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 when per-catalog views are unsupported.
Embeds Immuta documentation and an AI-powered chatbot directly in the Immuta UI, accessible from the top menu Documentation icon.
7 details in Everything
└──▷ BREAKING ON UPGRADE
!Immuta policy role names used to grant users access to Snowflake data sources will be renamed as part of the Snowflake integration architecture update rolling out beginning July 23, 2026; policy roles will be recreated under new names.
!Integration error banners for Snowflake integration validation failures will no longer be displayed; errors will only be available through the connections experience.
Cursor is an AI-powered code editor that helps developers write, edit, and debug code faster using integrated large language model assistance.
Cursor launches India-local Start plan at ₹649/month with UPI billing, cloud agents, and iOS remote control.
└──▷ HOW TO FIND IT
Sign up as a new user in India and select the Start plan during onboarding.
📍Visit cursor.com/signup and select the Start plan during onboarding to enroll at ₹649/month.
New Cursor Start plan (₹649/month, INR billing with UPI or card) targets developers in India with local pricing.
Includes always-on cloud agents that build, test, and ship code in the background while the developer keeps working.
Adds Cursor for iOS with remote control to launch and steer agents from a mobile device.
Bundles access to Grok 4.5 and Composer models alongside plugins, 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, hooks, and skills.
OpenCode v1.18.8 upgrades the 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 SDK to v2 and adds a project selector keyboard shortcut.
└──▷ GET THIS VERSION
$ git clone --branch v1.18.8 https://github.com/anomalyco/opencode.git
# already have the repo? check out this version:$ git checkout v1.18.8
Upgrades the MCP client SDK to v2, improving compatibility with newer MCP servers and OAuth flows.
Adds a mod+shift+o shortcut to open the project selector from the new-session page in the desktop app.
Lightweight coding agent that runs in your terminal
Codex CLI alpha.14 adds cloud-managed profiles for codex sandbox, network policy callbacks for remote exec, and configurable developer instructions for v2 subagents.
└──▷ GET THIS VERSION
$ git clone --branch rust-v0.146.0-alpha.14 https://github.com/openai/codex.git
# already have the repo? check out this version:$ git checkout rust-v0.146.0-alpha.14
Loads cloud-managed profiles for codex sandbox, enabling permission profiles and managed requirements to be sourced from cloud config bundles.
Adds configurable developer instructions for v2 subagents, allowing per-subagent instruction customization in multi-agent workflows.
Routes curated plugins by authentication mode, so plugin recommendations are matched to the active auth context.
Prepares 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. and plugin recommendations concurrently, reducing latency when loading tool suggestions.
Terminates Windows non-TTY processes on interrupt, routing Ctrl-C through the termination callback for pipe-backed processes.
$ git clone --branch desktop-v0.0.6 https://github.com/cline/cline.git
# already have the repo? check out this version:$ git checkout desktop-v0.0.6
Adds a queued-messages panel above the composer showing a count with expand controls to edit, send immediately, or delete individual queued turns.
Adds a persistent sidebar update indicator that displays the new version and a one-click restart button after an update is downloaded, surviving toast dismissal.
Displays the full workspace name and current git branch in the header and sidebar, with stable branch tracking that survives transient git lookups.
3 details in Everything
1 more release in this issue
· 2026-07-27 → 2026-07-28
Cline desktop v0.0.5 adds a Cline API key onboarding path, cancellable browser sign-in, and consistent 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 cards with install guidance.
└──▷ GET THIS VERSION
$ git clone --branch desktop-v0.0.5 https://github.com/cline/cline.git
# already have the repo? check out this version:$ git checkout desktop-v0.0.5
Adds a Cline API key path to onboarding so users can authenticate without a browser.
Enables cancellation of a pending browser sign-in instead of waiting indefinitely.
Unifies MCP server cards across marketplace views with a single uninstall action and setup guidance on installed servers.
Delivers a major performance overhaul: animated background locked at 60 fps, composer keystroke stutter reduced from 245 slow keystrokes to 3, streaming responses coalesce per-token re-renders, and app boot fetches the provider catalog once instead of three times.
Framework for orchestrating role-playing, autonomous AI agents. By fostering collaborative intelligence, CrewAI empowers agents to work together seamlessly, tackling complex tasks.
CrewAI 1.15.8 adds WaitTool to let agents pause execution while waiting on long-running async jobs.
└──▷ GET THIS VERSION
$ git clone --branch 1.15.8 https://github.com/crewAIInc/crewAI.git
# already have the repo? check out this version:$ git checkout 1.15.8
HolmesGPT 0.37.0 adds JSON log scraping, Langfuse/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 pre-acquired Azure AD token support.
└──▷ GET THIS VERSION
$ git clone --branch 0.37.0 https://github.com/HolmesGPT/holmesgpt.git
# already have the repo? check out this version:$ git checkout 0.37.0
Adds JSON logging format support for log scrapers, enabling structured log ingestion pipelines.
Adds Langfuse integration via OpenTelemetry tracing for 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. observability and trace export.
Supports pre-acquired AZURE_AD_TOKEN for Entra IDEntra IDMicrosoft's cloud-based identity and access management service, formerly Azure Active Directory, used by cyber tools to authenticate users and enforce access policies in Microsoft 365 and Azure environments. authentication, allowing token reuse without re-authentication.
vMLX 1.6.18 adds a process-wide SSD cache budget with LRU eviction and detailed cache reporting.
└──▷ GET THIS VERSION
$ git clone --branch v1.6.18 https://github.com/jjang-ai/vmlx.git
# already have the repo? check out this version:$ git checkout v1.6.18
Adds a process-wide SSD cache budget with coordinated LRU eviction and off-request-path publication for improved prefix reuse management.
Introduces detailed cache reporting covering attempted prefixes, applied reuse, uncached suffixes, reconstruction, fallback prefillprefillThe first phase of inference, where the model reads the whole prompt in parallel before producing its first token. It drives time-to-first-token; decode is the token-by-token phase after it., and aggregate disk usage.
Supports MiniMax tool arguments emitted as request-schema XML children.
Retains media across Responses tool continuations and enables local attachment playback in the Electron app.
An MCP that lets AI tools securely connect to your infrastructure, write IaaS code, debug issues, and assist during incidents - without risking production stability. Built for security teams to approve and infrastructure teams to experience like magic.
Emisar v0.35.0: hostname-based runner identity, retention safety for connected runners, and faster behavior matrix CI.
└──▷ GET THIS VERSION
$ git clone --branch v0.35.0 https://github.com/AndrewDryga/emisar.git
# already have the repo? check out this version:$ git checkout v0.35.0
Runners without an explicit runner.id now use their hostname as identity, so reboots reconnect the same runner and replaced ephemeral hosts auto-enroll without a persisted identity file.
Pack retention no longer sweeps versions still advertised by a connected runner, preventing pack_untrusted dispatch failures after idle-window cleanups.
Hosted admin runner bootstraps from a version-pinned release download with retries, eliminating GitHub API anonymous rate-limit failures on control-plane VM start.
Behavior cases for high/critical pack actions now run in disposable Compose projects as a non-root identity against real services, with network-path readiness probes enforced before cases start.
The snmp pack's OSPF adjacency fixture supports AppArmorAppArmorA Linux kernel security module, maintained as part of the kernel, that confines programs to a declared set of files, capabilities, and network access — letting cyber tools sandbox processes without writing custom kernel code.-confined hosts via network namespace construction.
+2 moreshow less
Behavior matrix CI runs in under half previous wall-clock time via pre-pulled images, sharded slow suites, and concurrency caps on heavyweight services.
CI now enforces that every high/critical action changed by the argument-boundary sweep has a successful behavior case or a declared machine-readable risk exception.
7 details in Everything
1 more release in this issue
· 2026-07-26 → 2026-07-27
emisar v0.34.0 adds streaming run output via 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. cursor paging and authoring-time pack argument validation.
└──▷ GET THIS VERSION
$ git clone --branch v0.34.0 https://github.com/AndrewDryga/emisar.git
# already have the repo? check out this version:$ git checkout v0.34.0
New wait_for_run MCP call accepts an output cursor and returns the next cursor, enabling agents to stream run output incrementally as it arrives rather than polling.
Curl-backed API actions now fail on 4xx/5xx HTTP responses instead of reporting transport success, making pack error handling reliable.
Pre-run dispatch rejections (contract changes, refusals, rate limits) are now logged with bounded, allowlisted fields, making refused MCP calls visible in operations.
Pack-trust conflict errors now name the specific runners that disagree about an action instead of failing generically.
Operator input (approval notes, grant scopes, policy overrides) now survives console re-renders triggered by co-approver broadcasts or refused submits.
+2 moreshow less
Added llms.txtllms.txtA proposed convention for a plain-text file placed at a website's root to give LLMs a concise, structured summary of the site's content, helping AI agents navigate it more reliably than parsing raw HTML. index and a Keycloak setup guide to documentation.
Documentation reorganized by task: SSOSSOSingle sign-on, using one identity provider to sign in to multiple applications. It reduces separate passwords and centralizes account and access management. and directory sync split, cloud and CLI agent connection split, containers separated into Kubernetes and Nomad sections with a host-install page, and operational limits and autoscaling fleets given dedicated pages.
Security scanner for AI agent skills. Detect vulnerabilities, malicious patterns, and security risks.
SkillSpector v2.5.0 adds inspection-ledger reporting with execution-completeness fields in JSON/SARIFSARIFStatic Analysis Results Interchange Format, a standard for sharing code-analysis findings, including locations and rule details. It lets scanners report results to a common review interface. and v2 baseline fingerprints.
└──▷ GET THIS VERSION
$ git clone --branch v2.5.0 https://github.com/NVIDIA/SkillSpector.git
# already have the repo? check out this version:$ git checkout v2.5.0
└──▷ TRY IT
Check whether a scan completed reliably in CI — block the pipeline if execution_successful is false rather than silently accepting a zero-finding report.
Adds canonical inspection-ledger accounting across static 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. analysis stages, including per-component coverage and explicit out-of-scope records.
New execution_successful top-level field and analysis_completeness.ledger_exceptions diagnostics in JSON output let automation block incomplete or failed scans.
JSON and SARIF reports now include execution-completeness information and analyzer status so a zero-finding report can be distinguished from a partial or failed scan.
Recursive scans now propagate child-scan failures to the combined report and return a failure when any child scan fails.
CLI now exits with code 2 for fatal execution or accounting failures, even when a JSON report was produced.
+1 moreshow less
Baseline fingerprints upgraded to version 2 format, binding accepted findings to the scanner version, source content, and full finding evidence.
6 details in Everything
└──▷ BREAKING ON UPGRADE
!Baseline files with version 1 fingerprints are no longer accepted; run skillspector baseline <path>, review the generated version 2 entries, and commit the replacement baseline (rules-only version 1 baselines remain supported with a warning).
!JSON integrations must treat invalid or missing output, a nonzero process exit, or execution_successful: false as a blocking validation error and surface analysis_completeness.ledger_exceptions for diagnosis.
THE CYBER TOOLCHAIN