AGENT-ARTIFACT-FIREWALL

Version updated for https://github.com/rustwoodagent-ops/agent-artifact-firewall to version v0.1.3.

  • 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-ARTIFACT-FIREWALL is a GitHub Action and CLI tool designed to scan and analyze AI agent-related artifacts—such as skills, plugins, hooks, and configuration files—for security risks before execution. It automates the identification of potential threats, including prompt injections, unsafe commands, and credential leaks, ensuring safer deployment of AI agents. Key features include customizable risk thresholds, various output formats (e.g., JSON, SARIF), and integration with CI pipelines for automated security checks.

What’s Changed

v0.1.3 makes AGENT-ARTIFACT-FIREWALL easier to try, easier to trust, and easier to demo.

What changed

  • upgraded the README to an activation-first flow
  • moved install and first-run output higher up the page
  • added deterministic local demo fixtures for:
    • AAF017 + AAF018
    • AAF015
    • AAF007 + AAF017
  • added end-to-end scan tests for those fixtures
  • refreshed the scanner test matrix doc

Why it matters

  • faster first successful runs in external repos
  • clearer proof of real scanner coverage
  • stronger public-facing demo story

Validation

  • deterministic fixture scans verified locally
  • go test ./internal/aaf passed
  • no scanner engine logic changed

Scope

  • README, docs, fixtures, and tests only
  • no rule expansion
  • no engine rewrite