#safeagents
1 articles tagged safeagents
Tagged articles
Articles tagged safeagents, newest first.

AI Security
safeagents Proxy Stops Coding Agents Tricked by Hidden README Instructions
Hidden README comment made a coding agent leak five .env secrets in 2 of 3 runs. Hugging Face's safeagents proxy inspects every tool call between agent and model. Three layers: allowlist checks, atif-scan pattern rules, optional LLM judge. Defaults missed the leak; a custom plugin of about 30 lines blocked every attempt. Pair the proxy with Docker sandboxing and tight permissions for unattended agents.
Get new articles by RSS or podcast
Add the RSS URL to your RSS reader or the podcast URL to your podcast app to get every new article as soon as it is published. The podcast reads out the headline and key points.
How to subscribe