This Week's Highlights — issue 020, November 14, 2026
The best tooling updates that shipped this week.
The week in numbers · Nov 7–13
Changes per week, last 12 weeks · Nov 7–13
Changes we read in that week's dailies, including docs and API changes. A gap means no daily was published.
Category moves this month
Every feature release from the 1325 tools on our watchlist goes into the daily newsletter. Once a week we read the whole field side by side and grade each release from 1 to 5 on three things, weighted equally:
- Depth. How complete is the best new capability?
- Impact. How much does it change what you can do?
- Novelty. Has any other tool done this before? We check every capability against the history of every tool we track. Features most tools already have earn nothing here. Being first earns the most.
A tool is judged on everything it shipped that week, so a project that releases daily gets credit for the sum — and still only takes one slot. We would rather run a short list than a padded one.
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.
This issue came out before editions were ranked, so it shows the Standard edition in issue order.
The picks in full
Semgrep is a static analysis tool that scans code for bugs, security vulnerabilities, and policy violations using customizable rules.
Semgrep ships Mandoline — a new program analysis engine scoring 100% on the OWASP Benchmark for Java with no AI token usage — alongside Guardian AI-powered detection of IDOR and BOLA at code-generation time, and a Context Layer that surfaces 2x–7x more real vulnerabilities in agentic workflows; together these close the gap between rule-based SAST and the class of business-logic flaws that rule engines structurally cannot find.
Semgrep shipped a hosted 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 that lets AI coding agents triage findings, run scans, and manage policies, alongside a major expansion of Semgrep Guardian — org-wide policy enforcement, AI-powered IDORIDORInsecure Direct Object Reference, a web vulnerability where an application exposes internal identifiers (IDs, filenames) without access checks, letting an attacker substitute another user's ID to reach their data./BOLABOLABroken Object Level Authorization, an API vulnerability class where a server fails to verify that the requesting user owns or may access the object ID they supply, letting attackers read or modify other users' data. detection, lockfile protection, and installs across Claude Code, Codex, and Copilot — plus new Supply Chain lockfile and registry coverage, broader language parsing, and several breaking changes to CLI platform support and access control.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
what's the number?
Each feature carries 0–100 for how completely the vendor documented it — not how big or important the work is. A major capability described in eight words scores low, and that is the finding.
- depth0–40 what it does and how it works or what changed
- specificity0–30 names real surfaces — APIs, flags, formats, limits, numbers
- actionability0–30 enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point
Code and config rank first by construction: a documented endpoint, flag or config key scores at the top of specificity and actionability, so it sorts above a feature described only in prose. Under 60 the notes go thin — everything below the “thinner coverage” line is thinner documentation, not smaller work. Hover any meter for that feature's three sub-scores.
Semgrep's hosted MCP server (public beta) lets Claude Code, Codex, Cursor, VS Code, ChatGPT, or any MCP client that supports remote HTTP servers search and filter findings across Semgrep Code, Supply Chain, Secrets, and Agentic Workflows, triage and reopen findings in bulk, read code snippets and finding history, start and retry managed scans, read scan logs, look up advisories and dependencies, dry-run and apply policy changes, manage Guardian rules, pull reports by severity/product/rule, trigger and schedule Agentic Workflows (admins only), create and link Jira/Linear tickets, and check SCMSCMSource Code Management: version-control systems (Git, SVN, etc.) that track changes to code over time, giving cyber tools reproducible builds, audit trails, and collaborative development workflows. connections — with permissions following your AppSecAppSecApplication security: the practice of finding and fixing vulnerabilities in software during design, development, and testing. Cyber tools build on AppSec workflows to shift vulnerability detection earlier in the development lifecycle. Platform role.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄ Authenticate via OAuth (browser sign-in scoped to an organization) or an API token: a Web API token allows all read/write operations, a Read-only token allows reads only, and an Agent (CI) token covers read/write on scans (Member tokens aren't supported). A discover tool finds the right MCP tool for a task and a semgrep_account tool returns the connected organization. The semgrep mcp server can also scan files outside the MCP client's workspace root by setting SEMGREP_MCP_ALLOW_SCAN_OUTSIDE_WORKSPACE=true. Available to organizations on Teams or Enterprise plans (trials included).
$ export SEMGREP_MCP_ALLOW_SCAN_OUTSIDE_WORKSPACE=true && semgrep mcp
discover tool to list available capabilities or semgrep_account to confirm which organization it is connected to.Security teams can define custom YAML rules (such as no-legacy-auth) in a local rules file or repo-wide policy file that Guardian automatically applies to every developer's agent session without manual distribution. Developers can also store local rules in ~/.semgrep/guardian-rules/ or the root of a repository; Guardian runs up to 1,000 organization rules and 50 local developer rules per scan. Guardian Policies and local developer rules are now generally available for all organizations, with per-rule control in the Claude Code remote plugin and the Codex plugin.
rules:
- id: no-legacy-auth
languages: [python]
severity: ERROR
message: Use auth.v2.verify() — legacy_auth is deprecated.
pattern: legacy_auth(...)
$ cp my-org-rules.yaml ~/.semgrep/guardian-rules/
Semgrep Guardian can now be installed directly into coding agents: in Claude Code via claude plugin install semgrep@claude-plugins-official, in Codex via codex plugin marketplace add https://github.com/semgrep/guardian --sparse .claude-plugin --sparse plugin followed by codex plugin add semgrep@semgrep-marketplace, and in Copilot via copilot plugin marketplace add https://github.com/semgrep/guardian followed by copilot plugin install guardian-copilot@semgrep-marketplace. Guardian also supports Microsoft Copilot alongside Codex and Claude Code, Codex now gets automatic code, dependency, and secret scanningsecret scanningChecking files, code or other data for exposed credentials such as passwords, API keys and tokens. Findings help teams remove the secret and revoke or rotate it. without a Semgrep CLI install, and Semgrep Guardian for GitHub Copilot CLI is in public beta on macOS and Linux.
$ claude plugin install semgrep@claude-plugins-official
$ codex plugin marketplace add https://github.com/semgrep/guardian --sparse .claude-plugin --sparse plugin codex plugin add semgrep@semgrep-marketplace
Supply Chain now shows the dependency path for transitive findings in Go projects when --allow-local-builds and dependency paths are turned on. With --allow-local-builds, running semgrep ci --no-suppress-errors now exits with code 2 when dependency resolution fails for a subproject, listing the affected dependency sources in scan output.
Set the SEMGREP_GUARDIAN_OBSERVE_ONLY environment variable to 1 to let Guardian scan as usual while hiding results from developers and coding agents until you're ready to roll it out. Observe Mode collects metrics in the background and makes no changes to the developer experience during staged rollouts, letting teams validate impact before enabling blocking.
$ export SEMGREP_GUARDIAN_OBSERVE_ONLY=1
The new Detect-2Shell Agentic Workflow finds Node.js and PHP versions across Docker images, .nvmrc files, infrastructure as code, and CI, marking each as vulnerable, unpinned, fixed, or unknown and using AI to resolve versions set in variables or base images. Agentic Workflows can now be scheduled to run daily, weekly, or monthly, which requires Unified Policies (the default for new organizations).
Supply Chain now parses bun.lock, giving Bun projects dependency, malware, and vulnerability coverage. Every package in requirements.txt, pinned or not, now appears in the dependency inventory and SBOMSBOMA software bill of materials, an inventory of the components and versions inside an application. It helps teams find affected software when a dependency has a vulnerability or licensing issue.. Supply Chain can also resolve private Maven or Gradle packages from AWS CodeArtifact for transitive dependency tracking, configured via an 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. role ARN under Settings > Supply Chain > Private Registries.
Agentic Workflows can now run on your own AI provider keys for OpenAI, Anthropic, Google Gemini, xAI, and Microsoft Foundry, so when configured all model requests and AI-powered detection use your providers instead of Semgrep-managed keys. Microsoft Foundry (formerly Azure OpenAI) gets its own card in AI provider settings accepting a resource endpoint, an API key, and one or more deployments each with its model. AI provider settings also now offer Automatic and Manual modes, Cloudflare Workers AI, and a direct connection to Anthropic with your own API key.
Semgrep Code now parses newer Bash arithmetic expressions such as $(( $X * 2 )), Clojure files using reader-evaluated metadata (^#=(...)) or the legacy #^ prefix, and Swift 5.9 syntax including macros, parameter packs, and if and switch expressions. It also parses C# Unicode escape sequences and Unicode identifier characters, Terraform provider functions with a namespace (e.g. provider::fn(...)), conditional and loop template directives, Lua 5.4 variable attributes, and Lua 5.5 leading attributes, global declarations, and named variable-argument parameters. Experimental support for Objective-C parses .m files and runs search-mode and taint-mode rules with metavariables and ellipsis operators.
Guardian blocks agent-initiated changes to lockfiles that would introduce malware or a reachable vulnerability, preventing unsafe dependency changes before they reach the build system. It also pushes advisories to every developer to check whether the version on their machine is still vulnerable when a critical supply chain incident (such as PHP2Shell or Node2Shell) is disclosed.
A new Advisories with findings filter on the Advisories page shows only advisories affecting your projects, respecting project-level 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. and excluding archived projects. An advisory's matching projects now include a 'Potentially matching: declared in a version range' section listing each project, branch, file, and declared version, and Incident Slack notifications report how many such projects exist and which packages they declare.
The Semgrep CLI no longer ships binaries for Intel Macs; version 1.182 is the last release supporting them. Use the semgrep/semgrep Docker image on Intel Mac or native arm64 Python on Apple silicon.
Findings CSV exports now include Last triaged by, Ticket key, and Ticket URL columns; ticket columns appear only when a Jira integration is active.
Autotriage cost drops to 0.1 AI credits per finding (down from 1 credit) starting October 31, 2026, for paid plans, trials, and the free plan.
Guardian's AI-powered detection reviews code at generation time and reasons about complex vulnerabilities including insecure direct object references (IDOR) and broken object-level authorization (BOLA), as well as code written with malicious intent.
Self-service organizations can now buy Semgrep Secrets standalone or bundled, and manage or cancel subscriptions via the Stripe customer portal by clicking Edit plan.
The issues API now returns an issueUrl field containing a link to each issue in AppSec Platform.
Adds a THIRD-PARTY LICENSES section to the --help output of most semgrep commands, linking to semgrep.dev/cli-credits.
The Wiz integration now sends findings from Azure DevOps, Bitbucket Cloud, and Bitbucket Data Center repositories in addition to GitHub and GitLab.
Users with the Read-only role and API tokens with the Read-only scope can no longer bulk ignore or reopen Semgrep Secrets findings; bulk triage now requires triage permission.
Free and Teams plan organizations can no longer select GitHub Enterprise, GitLab Self-Managed, or Bitbucket Data Center when setting up Semgrep Managed Scans.
New organizations can start a 30-day free trial of Guardian and upgrade in the product without a sales call.
Adds notes to Agentic Workflows issues via the New note button at the top of an issue's activity.
Guardian now alerts when a compromised version of the Claude Code extension for Cursor containing an SSH backdoor is detected.
Guardian updates itself automatically on new releases, with each download verified to SLSA Level 3SLSA Level 3A supply-chain integrity framework tier defined by the OpenSSF requiring that builds run on a hardened, ephemeral build platform and that provenance is generated and signed by the build service itself, not the developer..
Guardian hooks now run on PowerShell in addition to Bash, covering developers on Windows.
- New compliance framework Mandoline OWASP Benchmark and NIST Juliet test suite coverage
- New supported platform EU multi-tenant region
- New integration OpenAI Daybreak Defense Factory import integration
- !The Semgrep CLI no longer ships binaries for Intel Macs; version 1.182 is the last release supporting them. Use the `semgrep/semgrep` Docker image on Intel Mac or native arm64 Python on Apple silicon. Our best guesssnapshot-20261112
- !Users with the Read-only role and API tokens with the Read-only scope can no longer bulk ignore or reopen Semgrep Secrets findings; bulk triage now requires triage permission. Our best guesssnapshot-20261107
- !Free and Teams plan organizations can no longer select GitHub Enterprise, GitLab Self-Managed, or Bitbucket Data Center when setting up Semgrep Managed Scans. Our best guesssnapshot-20261107
Endor Labs is a software supply chain security platform that detects vulnerabilities and malicious dependencies in open source code.
Endor Labs AI SAST launches a multi-agent pipeline that orchestrates detection, triage, and remediation agents to find IDORs, broken access control, and insecure design across the full OWASP 2025 Top 10 — automatically classifying 95% of findings as false positives before they reach the security team — with C support delivering working exploits and call paths, and BYOK so model requests never leave the customer's own provider account.
Endor Labs launched AI SASTSASTStatic application security testing, analyzing source code or compiled code for vulnerabilities without running the application. It helps developers find problems before deployment. in private preview, a multi-agent system that triages findings, learns organizational context, and now covers C alongside bring-your-own-key model support.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
what's the number?
Each feature carries 0–100 for how completely the vendor documented it — not how big or important the work is. A major capability described in eight words scores low, and that is the finding.
- depth0–40 what it does and how it works or what changed
- specificity0–30 names real surfaces — APIs, flags, formats, limits, numbers
- actionability0–30 enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point
Code and config rank first by construction: a documented endpoint, flag or config key scores at the top of specificity and actionability, so it sorts above a feature described only in prose. Under 60 the notes go thin — everything below the “thinner coverage” line is thinner documentation, not smaller work. Hover any meter for that feature's three sub-scores.
Endor Labs AI SAST now automatically reads AGENTS.md, CLAUDE.md, skills files, and AI rules files from the repository to extract security-relevant context — such as authentication middleware and user-facing scope — without any configuration.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄ It also ingests 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. files — Dockerfiles, Terraform, 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., Kubernetes YAML, and Helm charts — to build a deployment model that infers network boundaries and component interactions for every PR and local developer scan. Users can upload architecture and threat model context in plain natural language, which the system verifies against code and classifies into three categories — project context, true positives, and false positives — scopeable to a single project, a set of projects, or the entire tenant. Reviewers can attach thumbs-up or thumbs-down feedback to individual findings during triage, teaching the scanner to adjust severity for entire classes of issues and suppress patterns the team has explicitly ruled on in future scans.
AI SAST and AI Chat model requests can now be routed through your own OpenAI, Azure OpenAI, Gemini, or Vertex AI account by enabling BYOKBYOKBring Your Own Key — a cloud deployment model where the customer supplies and controls the encryption keys rather than the provider, so sensitive data stays under the customer's cryptographic authority., with provider API keys either encrypted in the Endor Labs platform or passed as environment variables at scan time so keys are never stored with Endor Labs. BYOK enforces a provider allowlist and tenant isolation so requests go only to approved providers and your keys and requests remain separate from other tenants.
Endor Labs AI SAST now supports C with buildless scans that run in the IDE and on pull requests, covering buffer overflows, use-after-free, integer overflows, and cross-function flaws without requiring a compiler or build to reproduce. C SAST and C 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. now run together in the same platform, covering first-party C code and the open-source it depends on in a single pass. Each C finding includes a call path, a working exploit, and a suggested fix — the same evidence format already used for other supported languages.
Endor Labs launched AI SAST, in private preview, using multiple agents to detect business logic flaws and IDORs that rule-based tools miss. It automatically classifies every finding as true positive, false positive, or unknown across the full OWASPOWASPA nonprofit foundation that publishes open standards, guides, and ranked vulnerability lists—most famously the OWASP Top 10—giving cyber tools a vendor-neutral benchmark to test or report against. 2025 Top 10, including broken access control and insecure design, before results reach the security team. An AI Security Code Review agent assesses pull requests for architecture design and business logic flaws affecting security posture, working alongside the SAST detection and triage agents.
- New integration OpenAI Daybreak Defense Factory / Codex Security integration
- New integration OpenAI Daybreak Defense Factory / Codex Security integration
Zscaler's Zero Trust Browser now streams thick-client applications — SAP GUI, database utilities, developer IDEs — inside isolated, ephemeral cloud environments rendered directly in a browser tab, with no local VDI client, endpoint agent, or customer-managed infrastructure required; this closes the last major gap in browser-delivered zero-trust access that previously forced organizations to maintain VDI for non-web apps.
Zscaler's Zero Trust Browser can now stream thick-client applications such as SAP GUIGUIA graphical user interface, controls such as windows, buttons and menus. It lets people use software visually rather than entering commands. and developer IDEs from isolated, ephemeral cloud environments, eliminating the need for VDIVDIVirtual Desktop Infrastructure: a model where desktop operating systems run as VMs on a central server and stream to thin clients, giving cyber tools a single managed environment to monitor or scan..︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
what's the number?
Each feature carries 0–100 for how completely the vendor documented it — not how big or important the work is. A major capability described in eight words scores low, and that is the finding.
- depth0–40 what it does and how it works or what changed
- specificity0–30 names real surfaces — APIs, flags, formats, limits, numbers
- actionability0–30 enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point
Code and config rank first by construction: a documented endpoint, flag or config key scores at the top of specificity and actionability, so it sorts above a feature described only in prose. Under 60 the notes go thin — everything below the “thinner coverage” line is thinner documentation, not smaller work. Hover any meter for that feature's three sub-scores.
The Zero Trust Browser now supports thick-client application delivery through Application Isolation, running apps like SAP GUI, database utilities, and developer IDEs inside isolated, ephemeral cloud environments rendered directly in a browser tab, with no local VDI client, endpoint agent, or customer-managed infrastructure required.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄ Users access these thick-client desktop environments from the Zero Trust Browser landing page alongside existing web and private applications. Session persistence is enabled by default, letting users disconnect and reconnect with unsaved work intact, and copy/paste plus file downloads from isolated sessions are governed by policy to keep enterprise data inside the cloud environment.
Prowler adds an emerging threat signature engine — no prior art found — that lets practitioners supply threat signatures per scan to match compute instances across AWS, Azure, and GCP against affected image publishers, name patterns, version ranges, and exposed ports, with Amazon Inspector CVE correlation to automatically clear false positives; three new checks fire on matched instances and store structured threat data in a new Check_Report.raw payload.
Prowler shipped emerging-threat signature matching for AWS, Azure, and Google Cloud compute instances, real-time AWS detection and remediation in Prowler Cloud, and dozens of new checks spanning AWS Bedrock AgentCore, Cloudflare, GitHub, and Vercel.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
what's the number?
Each feature carries 0–100 for how completely the vendor documented it — not how big or important the work is. A major capability described in eight words scores low, and that is the finding.
- depth0–40 what it does and how it works or what changed
- specificity0–30 names real surfaces — APIs, flags, formats, limits, numbers
- actionability0–30 enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point
Code and config rank first by construction: a documented endpoint, flag or config key scores at the top of specificity and actionability, so it sorts above a feature described only in prose. Under 60 the notes go thin — everything below the “thinner coverage” line is thinner documentation, not smaller work. Hover any meter for that feature's three sub-scores.
Supply threat signatures per scan with the threat_signatures configuration key for AWS, Azure, and Google Cloud to match compute instances against affected image publishers, name patterns, per-branch version ranges, exposed ports, impact, and remediation.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄ Three new checks — ec2_instance_emerging_threat_image (AWS), vm_emerging_threat_image (Azure), and compute_instance_emerging_threat_image (Google Cloud) — flag instances running images matched by a loaded threat signature, with matched-threat data attached to each finding via the new Check_Report.raw payload. On AWS, the engine clears a match when Amazon Inspector finds no active CVE finding for the threat, using the new inspector2:SearchVulnerabilities permission added to the Prowler additions policy and the 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. scan role. In Prowler Cloud, issues can be filtered by threat with filter[threat] or searched by threat name/CVE ID with filter[search] on GET /api/v1/issues and GET /api/v1/issues/overview.
$ curl -s -H 'Authorization: Bearer <token>' 'https://api.prowler.com/api/v1/issues?filter[threat]=my-threat-001'
Prowler Cloud ingests CloudTrailCloudTrailAWS service that records API calls and account activity across an AWS environment, giving cyber tools an audit log to detect anomalous or malicious actions. events via EventBridgeEventBridgeAWS's managed event bus service that routes events between cloud services, SaaS apps, and custom code using rules and targets, letting tools react to state changes without polling or managing message infrastructure./SQS or a LookupEvents fallback to detect AWS activity in real time, with per-account policy modes monitor, pr, and auto (including dry-run support), a new /realtime dashboard, and a Manage Realtime role permission to control who can change the real-time policy. Lighthouse AI and the MCP ServerMCP ServerA process that exposes tools, resources, and prompts over Anthropic's Model Context Protocol, letting an AI agent call external APIs or local services through a standardised interface. gain tools to manage the event feed, read detections and actions, and revert them.
AWS gains a bedrockagentcore service with bedrockagentcore_full_access_policy_attached, bedrockagentcore_payments_process_payment_role_separation, bedrockagentcore_runtime_vpc_configured, plus cloudtrail_agentcore_data_events_enabled for AgentCore data event coverage. Cloudflare gains 13 checks: a new account service (account_two_factor_enforced, account_member_two_factor_enabled, account_super_admins_limited, account_api_token_expiration_set, account_api_token_ip_restriction_enabled, account_api_token_recently_used), a new r2 service (r2_bucket_public_dev_url_disabled, r2_bucket_cors_no_wildcard_origin, r2_bucket_custom_domain_min_tls_secure), a new workers service (workers_script_observability_enabled, workers_script_workers_dev_disabled, workers_script_preview_urls_disabled, workers_script_no_secrets_in_plain_text_bindings), plus zone_firewall_no_skip_all_traffic_rules. GitHub gains repository_secret_scanning_push_protection_enabled to flag repositories that accept pushes containing secrets. Vercel gains project_environment_no_secrets_with_public_prefix (flagging secret-like environment variables exposed via prefixes such as NEXT_PUBLIC_ or VITE_) and project_linked_to_git_repository (flagging projects that can only receive unreviewed CLI deployments).
Set PROWLER_NO_PROMOTIONS=1 to suppress the Prowler Cloud hand-off note in the OSSOSSOpen-source software: code whose licence permits anyone to inspect, modify, and redistribute it. Cyber tools build on it to reuse auditable libraries without licensing fees. CLI dashboard; Prowler Cloud and Private Cloud reports never show it.
$ PROWLER_NO_PROMOTIONS=1 prowler dashboard
New check oam_sink_not_publicly_accessible reports CloudWatch Observability Access Manager sinks whose resource policy allows any AWS account to link to them. New check organizations_security_services_delegated_admin_not_management_account verifies that security services with trusted access in the organization use a delegated administrator rather than the management account.
Prowler adds a 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. framework for Azure mapping Articles 25, 30, and 32 to existing Azure checks, and a Cloudflare mapping for NCSC Cyber EssentialsCyber EssentialsA UK government-backed certification scheme that defines a baseline set of technical controls organizations must implement to defend against common cyber threats, giving cyber tools a compliance benchmark to assess or report against. 3.3.
A Contextual Fix Lighthouse AI skill on the issue side panel reads open attack-path legs and findings and leads with the single change that closes the issue fastest.
The Findings page now shows findings from a running first scan as they arrive, with a View Early Findings action on executing scans; critical and high findings are stored first.
The issue side panel now holds everything the full issue page had — Findings and Resources tabs, leg evidence, compliance frameworks, and Send to Jira — and the separate issue page is removed.
Prowler Cloud shows labelled sample data from a demo AWS account on Overview, Findings, Resources, Compliance, Issues, and Attack Paths until the first scan completes.
Prowler Private Cloud is now available on AWS Marketplace as Pay as You Go and 10- and 25-account bundles, installed in the customer's own AWS account.
Prowler Cloud's Lighthouse AI settings now accept a Bedrock API key as an alternative to AWS access keys.
Teleport Beams runs AI agents in micro-VMs with delegated X.509 identities scoped to the current task (auto-expiring in 12 hours), MITRE ATT&CK session scoring, Phases for Git-like filesystem snapshot and fork across runs, agent-initiated JIT access requests, and MCP tool-level RBAC with short-lived certificates — the first tracked tool to give AI agents hardware-isolated, identity-scoped, auditable infrastructure access rather than inheriting a user's full role set.
Teleport's Beams runtime gained "Phases," a Git-like system for managing container filesystem state with snapshot, fork, rebase, and rebuild workflows, plus a way to run persistent background agent tasks inside a beam.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄
what's the number?
Each feature carries 0–100 for how completely the vendor documented it — not how big or important the work is. A major capability described in eight words scores low, and that is the finding.
- depth0–40 what it does and how it works or what changed
- specificity0–30 names real surfaces — APIs, flags, formats, limits, numbers
- actionability0–30 enough to go use it — a named endpoint, flag, or config key tops this; a UI path is a starting point
Code and config rank first by construction: a documented endpoint, flag or config key scores at the top of specificity and actionability, so it sorts above a feature described only in prose. Under 60 the notes go thin — everything below the “thinner coverage” line is thinner documentation, not smaller work. Hover any meter for that feature's three sub-scores.
Beams runtime adds Phases, defined in a Kube-style phase.yaml file supporting version, metadata.name, from, files, commands, and services fields, giving snapshot, fork, rebase, and periodic rebuild capabilities for a container's filesystem.︀︄︇︀︅︉︀︅︀︀︃︀︀︅︅︀︅︊︀︄︆︀︃︄ Create a phase interactively with tsh phase add <name>, fork an existing beam's filesystem state with tsh phase add <name> --fork, apply a definition with tsh phase apply -f phase.yaml, and save or finalize the current /home state as a durable snapshot with tsh phase snapshot or tsh phase snapshot --final. The services block can start a pre-start or background service inside a phase via type: pre-start or type: background, scoped to context: user or context: snapshot.
$ tsh phase add beak-buttons --fork beamctl start sh -c 'codex exec -C ~/parakeets "Add a beak-friendly interface. Test it, commit, push, then run tsh phase snapshot." </dev/null'
version: v1
metadata:
name: dev-files
files:
- from: ./dev-files
to: /opt/dev
---
version: v1
metadata:
name: environment
from: dev-files
commands:
- shell: |
apt-get update
apt-get install -y curl
- shell: |
v=$(curl -fsSL 'https://go.dev/dl/?mode=json' | jq -r '.[].version | select(startswith("go1.27."))')
curl -fsSL "https://go.dev/dl/$v.linux-amd64.tar.gz" | tar -xz -C /usr/local
ln -sf /usr/local/go/bin/go /usr/local/bin/go
services:
- name: dev-server
type: background
context: snapshot
shell: |
export GOCACHE=/tmp/go-build GOPATH=/tmp/go
go run /home/beams/parakeets/main.go
$ tsh phase apply -f phase.yaml tsh phase add dev --from=environment
Run a persistent agent task inside a beam without staying connected using beamctl start <command>, shown forking a beam and kicking off a long-running agent command that commits and snapshots on its own.
$ tsh phase add beak-buttons --fork beamctl start sh -c 'codex exec -C ~/parakeets "Add a beak-friendly interface. Test it, commit, push, then run tsh phase snapshot." </dev/null'
- New detection MITRE ATT&CK session scoring for AI agent behavior
Where each pick stands · Nov 7–13
Tools with changes in each pick's category over the last 12 weeks, including the pick itself.
Each pick's release rhythm · Nov 7–13
Changes per week for the last 12 weeks. The bright bar is this week; a dash marks a week with no daily.
Which categories moved · Nov 7–13
The thick bar is this week. The thin bar is the average of the four weeks before, counting only weeks with a daily.
Who shipped the most · Nov 7–13
Changes per tool this week. Volume is not a quality score.
Calls for papers, tickets, dates, talks and recordings this week. Star a conference to hear about the ones you follow.