Aegis AI Governance Gate

Version updated for https://github.com/Acacian/aegis to version v0.9.2.

  • This action is used across all versions by ? repositories.

Action Type

This is a Composite action.

Go to the GitHub Marketplace to find the latest changes.

Action Summary

Agent-Aegis is a GitHub Action designed to identify and mitigate ungoverned AI and tool calls in your codebase, helping developers ensure compliance with governance policies before deploying to production. It scans code to detect risky patterns across multiple frameworks, such as unregulated API calls, data leakage, and uncontrolled code execution, and provides automated solutions to fix these issues. By adding a single line of code, it enables features like policy enforcement, PII masking, injection blocking, and audit trail generation without requiring manual code modifications.

What’s Changed

What’s New

Scan-First Experience

  • README restructured around aegis scan — “Find ungoverned AI calls in 30 seconds” is now the first thing you see
  • Playground: new aegis scan demo tab as the default landing experience with 4 presets

GitHub Action: PR Comments

  • aegis GitHub Action now posts policy results (scan/score/plan/test) as PR comments
  • Upserts existing comments to avoid spam
  • New inputs: comment (true/false), github-token

Playground: Policy CI/CD Demo

  • New interactive “Policy CI/CD” tab with 4 scenarios
  • Simulates aegis plan → aegis test → PR comment preview workflow

Fixes & Cleanup

  • Windows compatibility: UTF-8 encoding for YAML files, path normalization, timer resolution
  • Removed dead MANIFEST.in (hatchling build)
  • Synced server.json version
  • Exported 5 previously unreachable modules (killswitch, autopolicy_llm, rate_limiter_redis, tenant, tiers)

Full Changelog: https://github.com/Acacian/aegis/compare/v0.9.1...v0.9.2