Automated output vetting for AI agents. A Hermes plugin that pre-screens what leaves the agent: suppresses small counts in tables (statistical disclosure control), redacts credentials, and holds data uploads for approval. Deterministic rules, policy file, masked audit log.
Stars, forks and merged pull requests each day or month, lined up with the repository's rank on Trendshift Trending
“AI agents leak data by accident: a tiny count in a table, a file uploaded to share, a key in a command. Egress Guard brings automated output vetting to Hermes agents, enforced in code, even on a…
@mornek315 · x.com
