AI agent tools for AI safety
Short answer: Our top pick for AI safety: agent-governance-toolkit, with the highest AgentPick Score (72/100) among tools tagged "ai-safety". Runners-up: cc-safety-net (70) and h5i (69). Compare what each does, its license and how recently it was updated below.
| # | Tool | What it does | Score | Stars | License | Updated |
|---|---|---|---|---|---|---|
| #1 | agent-governance-toolkit | AI Agent Governance Toolkit — Policy enforcement, zero-trust identity, execution sandboxing, and reliability engineering for autonomous AI agents. Cov… | 72/100 | ★ 6.4k | MIT | 14 Oct 2026 |
| #2 | cc-safety-net | A pre-execution guard for AI coding agents. It blocks destructive Git and file system commands, plus common attempts to access sensitive files, before… | 70/100 | ★ 1.6k | MIT | 14 Oct 2026 |
| #3 | h5i | An agent-native web security workspace. Find vulnerabilities through red-teaming. Formally verify properties of Rust codes in Lean 4. | 69/100 | ★ 777 | Apache-2.0 | 14 Oct 2026 |
| #4 | guardana | Open-source AI security verification for model artifacts, live endpoints, MCP servers, and recorded agent traces. Reproducible evidence for release de… | 63/100 | ★ 287 | Apache-2.0 | 14 Oct 2026 |
| #5 | node9-proxy | Access control for AI agents. Set what Claude Code, Codex, Gemini, Cursor and any MCP server are allowed to do, review risky actions before they run, … | 61/100 | ★ 220 | Apache-2.0 | 14 Oct 2026 |
| #6 | box | Run AI agents in a sandbox that restricts what they can execute, read, write, and reach on the network. Box combines OS isolation with default-deny Do… | 56/100 | ★ 403 | Apache-2.0 | 14 Oct 2026 |
| #7 | cordum | The action firewall for AI agents. Enforce policy and human approval before risky tool calls, shell commands, workflows, and production changes, with … | 52/100 | ★ 509 | See repo | 14 Oct 2026 |
| #8 | Aegis | Runtime policy enforcement for AI agents. Cryptographic audit trail, human-in-the-loop approvals, kill switch. Zero code changes. | 51/100 | ★ 542 | MIT | 9 Oct 2026 |
| #9 | PostTrainBench | Measuring how well CLI agents like Claude Code or Codex CLI can post-train base LLMs on a single H100 GPU in 10 hours | 38/100 | ★ 593 | MIT | 2 Oct 2026 |
| #10 | AgentLens | Agentlens is a trusted agent trading platform. Here, you can quickly find the Agent that meets your needs, and you can also publish your own Agent to … | 31/100 | ★ 1k | AGPL-3.0 | 24 Aug 2026 |
| #11 | rust-docs-mcp-server | 🦀 Prevents outdated Rust code suggestions from AI assistants. This MCP server fetches current crate docs, uses embeddings/LLMs, and provides accurate… | 17/100 | ★ 299 | MIT | 24 Nov 2025 |
Source: each project's public GitHub repository (description, stars, license, last push). Listed tools aren't tested or ranked by AgentPick.
How to choose
- Grant the narrowest permissions the server allows.
- Check whether it runs locally or calls a remote service.
- Prefer servers listed in the official MCP registry.
Get the best new agent tools every week. Free, with a verdict on each. Get the weekly digest
Related
AI agent tools for local-first setups →
MCP servers
AI agent tools for knowledge graphs and bases →
MCP servers
AI agent tools for RAG (retrieval-augmented generation) →
MCP servers
AI agent tools for MCP client →
MCP servers
AI agent tools for cybersecurity →
MCP servers
AI agent tools for DevOps →
MCP servers
AgentPick Weekly
The week's new agent tools with a verdict on each, every week. Free. We'll email you a link to confirm, and you can unsubscribe any time. See our Privacy Policy.