huggingface/safeagents
App
Can I trust it?
No red flags in our checks.
Stars from real accounts No data
We haven't checked its stars for fakes yet.
Active OK
Last change yesterday.
Licence OK
Apache-2.0: you can use it for free, also in a commercial product. Keep the licence notice.
Quick checks from public data. They help, but they are no guarantee.
What it is
A local proxy placed between a coding agent and its model server. It checks every tool call the model returns, blocks undeclared tools and suspicious calls via atif-scan detectors and an optional judge model, and replaces blocked calls with a text notice. Works with any harness…
Who it's for
Developers running AI coding agents who want to monitor or block risky tool calls before they execute
How to use it
Install
curl -LsSf https://raw.githubusercontent.com/huggingface/safeagents/main/scripts/install.sh | shThis command downloads a script and runs it, or needs admin rights. Review it before running it.
Ready-made installers, when the project publishes them, are on its latest release.
Ask your AI to set it up
Paste this into Claude, ChatGPT or your coding assistant.
Help me install and set up huggingface/safeagents in my project. Follow its README: https://github.com/huggingface/safeagents. Explain each step and check it is safe before running anything.
Star curve
Get alerts for repos like this
We'll email you when something in AI agents starts rising. Free, once a week, only what matches.
What arrives: one email on Mondays with the repos that match, never more. See last week's. Privacy
For maintainers
Add this badge to your README. It shows the repo's place on the radar, updates every hour and links back to this page.
[](https://repoalerts.com/repo/huggingface/safeagents?ref=badge)Similar repos
- FailproofAI/failproofai
Observability and enforcement layer for AI agent harnesses such as Claude Code, Codex, Hermes and OpenClaw. Records every agent run locally, provides a dashboard, and blocks dangerous tool calls befor
- IsuruMaduranga/harness-engineering-101
An open book series in Markdown that explains how AI agents work by building a roughly 300-line toy agent harness in plain Python. It covers the message array, tool calls, the agent loop, context mana
- vercel-labs/durable-harness
Runs coding-agent harnesses (Pi, OpenCode, fx, Claude Code, Codex) as durable agents on a Workflow SDK World. Each session is a workflow run whose conversation, turns and tool calls are kept in an eve
- collinzrj/claude-code-auto-safeguard-prompt
A system prompt and rule set for a security monitor that evaluates the latest action of autonomous AI coding agents and decides whether to block it. It defines hard and soft block categories covering
- Cygnux-Labs/Tracekit
Tamper-evident tracing and policy gating for AI agents. It checks each tool and model call before it runs, signs calls and decisions into a hash chain, and writes one verifiable file per run that can
- pzs19/HarnessPro
A framework that recursively improves the harness of an AI agent (its instructions, tools and runtime policies) by diagnosing execution traces and selecting candidate changes with synthesized tests. O