DAILY DROP · NOV 14, 2026

Today's drop: 2 new tools, 3 listing updates

The directory lists 150 hand-checked tools. New today: Vigilo, the AI pentest and code-review harness CERT-EU built for its own work and open-sourced on Nov 13, and Qwen Code, Alibaba's open-source coding agent for the terminal, now in a 1.0 Preview. Updated: Claude Code (v2.1.325, security fix), OpenAI Codex (CLI 0.169.0) and CISO Assistant (v4.3.0). AI for IT & Cybersecurity grows to 62 tools; we retired two listings to stay at 150. Pricing only where the official page states it.

Previous drop: Nov 13, 2026 →

Get the weekly dropThe best new AI tools, one email a week. Free.

New in the directory 2

Browse all 150
Tool
Category
Price
Status
Link
Vigilo (CERT-EU)Open-source AI penetration-testing and code-review harness built by CERT-EU, the cybersecurity service for the EU institutions, and open-sourced on Nov 13, 2026. Point it at a web application you own: specialised agents, one per vulnerability class (injection, XSS, authentication, SSRF, authorization, GraphQL, WebSocket, cryptography and supply chain), map the target, share findings through a common knowledge base and try to exploit them with captured evidence, then an adversarial critic re-checks every finding, challenges its severity and merges duplicates before it reaches the report. White-box mode reads the source (plus an optional running instance) and proposes a fix branch per confirmed finding; grey-box mode tests only the running app with valid credentials and adds detection hints for the blue team. Runs as a multi-user web platform (Docker Compose, optional GitLab integration) or a command-line pipeline, and drives an agentic coding CLI (Claude Code, OpenAI Codex or OpenCode) with cloud or on-premises models, so you supply the model access. CERT-EU says findings and patches must be validated by the code owners, and it was part of an ENISA proof of concept on AI source-code testing. EUPL-1.2 licence; new on GitHub (only a handful of stars so far). Only test systems you own or are authorized to test.
Price Free and open source (EUPL-1.2); you pay for the model access it uses
FREE TIERchecked 11/14
Qwen CodeOpen-source (Apache-2.0) AI coding agent from Alibaba's Qwen team for the terminal, with IDE plugins (VS Code, Zed, JetBrains), a desktop app, a web UI, SDKs and chat integrations. It has memory, skills, subagents, agent teams and MCP support, and talks to OpenAI, Anthropic, Gemini and Qwen APIs, any compatible third-party provider or a local model in Ollama or vLLM, switchable at runtime. Sign in with a Qwen account or bring your own API key. Install with the standalone script, npm (Node.js 22+) or Homebrew. Qwen Code 1.0 Preview (tagged Nov 11, 2026; preview.1 on Nov 12) merges the agent harness of Alibaba's Qoder with Qwen Code and ships native packages for macOS, Linux and Windows with no separate Node.js install; interfaces and settings may still change during the preview. The stable line is v0.27.3 (Nov 13). About 30,100 GitHub stars.
Price Free and open source (Apache-2.0); bring your own API key or use a Qwen account
FREE TIERchecked 11/14

Why these two

  • Vigilo is the harness CERT-EU, the cybersecurity service for the EU institutions, built to point AI agents at code from a security angle, and it open-sourced it on Nov 13. Instead of one agent, it runs specialised agents per vulnerability class (injection, XSS, authentication, SSRF, authorization, GraphQL, WebSocket, cryptography and supply chain) that share a knowledge base and try to prove each finding with evidence. Then an adversarial critic re-checks every finding, pushes back on inflated severities and merges duplicates, so the report is not a pile of maybes. With source code (white-box) it proposes a fix branch per confirmed finding; without it (grey-box) it tests the running app with valid credentials and adds detection hints for defenders. It drives Claude Code, Codex or OpenCode with cloud or on-premises models, so you bring the model access. It is brand new on GitHub with very few stars, so treat it as a serious reference design from a real CERT rather than a proven community tool, and only run it against apps you own, such as an intentionally vulnerable app in your home lab. CERT-EU itself says findings and patches must be checked by the code owners.
  • Qwen Code is one of the most used open-source coding agents (about 30,100 GitHub stars) and we had not listed it. It works with OpenAI, Anthropic, Gemini and Qwen APIs or a local model in Ollama or vLLM, which makes it a good no-lock-in option next to Claude Code and Codex. The 1.0 Preview (Nov 11) merges in the agent harness from Alibaba's Qoder and ships native installers for macOS, Linux and Windows; it is a preview, so settings may change. The stable v0.27.3 shipped Nov 13.

Listing updates

Checked on Nov 14, 2026 against official pages and release notes.

  • Claude Code: v2.1.325 (Nov 13) fixes a security issue in how zstd-compressed responses are decoded. It also tells you how to continue (run /rewind and change the message, or /clear) when your organization's policy blocks a request, and fixes /security-review stopping with an error in worktrees of bare repositories and --separate-git-dir checkouts. Update if you are on v2.1.323 or older.
  • OpenAI Codex: CLI 0.169.0 (Nov 13) keeps project configuration and MCP servers disabled for codex exec in untrusted directories, even with --sandbox workspace-write; pass --trust to codex exec or codex review to save project trust on purpose. You can now add or edit custom model providers from the terminal, the opt-in terminal workspace gets multiple side-chat tabs, and Windows sandbox handling for quoted cmd.exe paths is fixed.
  • CISO Assistant: v4.3.0 (Nov 13) adds service accounts federated with an external OIDC identity provider, an MCP tool for reviewing evidence, a more actionable landing-page briefing, a Hungarian NIS2 compliance preset and requirement names for the NIS2 2024/2690 framework, and tightens bulk-action permissions.

Category changes

Checked, not listed

  • BLA (Blind LLM Agents): a fast-rising MIT tool that adds decoys and integrity checks to protect your own code and model files from AI reverse-engineering agents. About 5,600 stars in under three weeks from a small team; we are watching whether the protection claims hold up.
  • Kitesurf (Cloudflare): the stateless browser for AI agents on Workers is now open source (Apache-2.0); still marked experimental.
  • Vaikora Guard MCP (Data443): MCP policy proxy, now MIT on PyPI; still watching traction.
  • OpenBot v1.0.3, 0 v0.28.11 and Hermes Agent canaries: routine releases, no listing change. Ollama v0.40.17 and Wazuh 4.14.11 are still release candidates.