v1.0.3

What's Changed

• Update default cagent version in the README.md by @derekmisler in #15
• feat: add macOS and Windows GHA runners support by @gabolaev in #16

New Contributors

• @gabolaev made their first contribution in #16

Full Changelog: v1.0.2...v1.0.3

v1.0.2

What's Changed

• Added a GH workflow as an example by @derekmisler in #12
  • Automatically generate a PR description with a /describe comment
• Update cagent version and replace 'cagent run --tui=false' with 'cagent exec' (which disables the tui by default) by @derekmisler in #11

Full Changelog: v1.0.1...v1.0.2

v1.0.1

What's Changed

• Add a GH workflow as an example by @derekmisler in #12
• Update cagent version and replace 'run --tui=false' with 'exec' by @derekmisler in #11

Full Changelog: v1.0.0...v1.0.1

🎉 cagent-action v1.0.0 - Initial Public Release

We're excited to announce the first public release of cagent-action, a GitHub Action for running CAgent AI agents in your CI/CD workflows with enterprise-grade security.

🚀 What is cagent-action?

Run AI agents directly in your GitHub workflows to automate code reviews, security scanning, documentation generation, and more. Built on CAgent's multi-agent architecture with comprehensive prompt injection protection.

✨ Key Features

Agent Execution

• Download & Run: Automatically fetches CAgent binary and executes agents from Docker Hub or local paths
• Multi-Provider Support: Works with Anthropic (Claude), OpenAI (GPT-4), Google (Gemini)
• MCP Gateway Integration: Optional containerized MCP server support for extended tool capabilities
• Interactive TUI Mode: Test agents locally with interactive approval prompts
• YOLO Mode: Auto-approve all agent actions for fully automated workflows
• Binary Caching: Smart caching reduces workflow execution time

Security Hardening

• Input Sanitization: Detects and blocks 45 HIGH-RISK prompt injection patterns
  • Instruction override attempts ("ignore previous instructions")
  • Direct secret extraction (echo $API_KEY, console.log(process.env))
  • System prompt manipulation
  • Jailbreak attempts
  • Encoding/obfuscation (base64, hex)
• Comment Removal: Strips code comments from PR diffs to prevent hidden injection vectors
• Output Scanning: Scans all agent responses for leaked API keys and tokens
• Automatic Incident Response: Creates security issues and fails workflows when threats detected
• Authorization Checks: Validates PR author permissions before execution

Developer Experience

• Simple YAML Configuration: Easy-to-use action inputs
• Comprehensive Outputs: Access exit codes, output files, and security flags
• Extensive Testing: 13 security tests + 6 exploit simulation tests
• Complete Documentation: Security guides, examples, and best practices

📦 Installation

Add to your workflow:

- name: Run AI Agent
  uses: docker/cagent-action@v1.0.0
  with:
    agent: agentcatalog/code-reviewer
    prompt: "Review these changes for bugs and security issues"
  env:
    ANTHROPIC_API_KEY: ${{ secrets.ANTHROPIC_API_KEY }}

🔒 Security Features

This action is designed with security-first principles:

Layer	Protection	Action
Input	45 HIGH-RISK patterns detected	Blocks execution
Input	4 MEDIUM-RISK patterns detected	Warns but continues
Execution	Isolated CAgent runtime	Sandboxed environment
Output	Secret pattern scanning	Blocks & creates incident

See our complete security documentation here.

📝 Example Workflows

Code Review Agent

name: AI Code Review
on: pull_request

jobs:
  review:
    runs-on: ubuntu-latest
    permissions:
      contents: read
      pull-requests: write
    steps:
      - uses: docker/cagent-action@v1.0.0
        with:
          agent: agentcatalog/code-reviewer
          prompt: "Review this PR for bugs, security issues, and best practices"
        env:
          ANTHROPIC_API_KEY: ${{ secrets.ANTHROPIC_API_KEY }}

Security Scanner

- uses: docker/cagent-action@v1.0.0
  with:
    agent: docker/github-action-security-scanner
    prompt: "Scan for vulnerabilities and security issues"
  env:
    OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}

Documentation Generator

- uses: docker/cagent-action@v1.0.0
  with:
    agent: agentcatalog/doc-generator
    prompt: "Generate API documentation from source code"
  env:
    ANTHROPIC_API_KEY: ${{ secrets.ANTHROPIC_API_KEY }}

📚 Documentation

• Getting started guide
• Security architecture and testing
• Example workflow configurations
• Contribution guidelines

🧪 Testing

All releases pass comprehensive security testing:

• ✅ 13 security hardening tests
• ✅ 6 exploit simulation tests
• ✅ Authorization and authentication checks
• ✅ Secret leak detection validation

Run tests locally:

cd tests
./test-security.sh
./test-exploits.sh

🙏 Contributors

Special thanks to the original contributors:

• Derek Misler
• Jean-Laurent de Morlhon

🐛 Known Issues

None at this time. Please submit an issue if you encounter any problems.

📖 Resources

• Learn more about cagent.
• Browse our agent catalog at Docker Hub.

🚦 What's Next?

We're actively working on:

• Additional pre-built agents for common workflows
• Enhanced MCP server integrations
• Performance optimizations
• Extended language model support