agent-security
https://github.com/solongate/agent-security
Go
Local-first, open-source security gateway for AI coding agents. Enforce tool-call policies, detect secrets, and audit actions on-device.
Triage Issues!
When you volunteer to triage issues, you'll receive an email each day with a link to an open issue that needs help in this project. You'll also receive instructions on how to triage issues.
Triage Docs!
Receive a documented method or class from your favorite GitHub repos in your inbox every day. If you're really pro, receive undocumented methods or classes and supercharge your commit history.
Go not yet supported1 Subscribers
Add a CodeTriage badge to agent-security
Help out
- Issues
- shared/paths: GuessPermission lets EXECUTE substrings shadow explicit WRITE keywords
- shared/dlp+state: custom DLP "regex" field is actually a restrictive glob, so real regexes silently never match
- shared/beat: WriteBeat/ReadBeat/BeatDir are dead code, duplicated (and diverged) elsewhere
- tui: App.remount() is dead code, so the GenMsg staleness guard never fires
- ratelimit: compact() never runs while an agent is being rate-limited, letting the log grow without bound
- dlp: EgressSecretCheck and RedactReadPlan both leave DLP coverage gaps that let secrets through
- guardcore: policy.json signature check fails open on any verification error or missing signature
- shield: shieldRedactBody misses array-form messages, non-message fields, and case-varied Content-Type
- shield: signal-relay goroutine leaks and proxied responses are never flushed
- tokens: readAntigravity builds a SQLite path from unsanitized conversationID, enabling path traversal
- Docs
- Go not yet supported