--- name: strix description: Autonomous AI penetration testing skill that finds and fixes vulnerabilities in your own applications and web services using the Strix AI pentest agent (usestrix/strix). Owner-authorized, defensive security auditing only. --- # Strix AI Penetration Testing ## Overview This skill integrates **[Strix](https://github.com/usestrix/strix)** — the open-source AI penetration testing tool — into the REA_Skill agent ecosystem. Strix deploys autonomous AI hackers to find and fix your app's vulnerabilities, with seamless GitHub Actions / CI/CD integration to scan on every pull request and block insecure code before it reaches production. Strix is **strictly for targets you own or have explicit, written authorization to test**. Unauthorized testing is illegal in most jurisdictions. ## What Strix Does - **Autonomous AI hacking:** AI agents that actively test your targets and propose fixes. - **CI/CD ready:** Integrates with GitHub Actions to scan every pull request. - **Cloud scans:** `strix cloud scans start --engagement-type live_test --domain-ids --wait` - **Vulnerability listing:** `strix cloud vulns list --severity critical` (JSON output supported) - **MCP servers:** Connect your own Model Context Protocol servers via `~/.strix/mcp-servers.json` and expose their tools to the agent during a run. - **REST API:** Every cloud operation has a matching `strix cloud ` command. ## ⚡ Integration with REA (Reverse Engineer Anything) When combined with the REA suite: 1. **Pre-publish vulnerability hunt:** Before shipping your app/website, run Strix to find weaknesses, then use REA skills to reverse-engineer and remediate the affected components. 2. **Web audit pivot:** Use the Website Security Audit (`website-security-audit`) skill for the 10-layer defensive scan, then Strix for active AI-driven penetration testing. 3. **Binary ↔ App correlation:** Recover source/assets with REA core skills, then point Strix at the running application to validate exploitability end-to-end. ```bash # Example: start a live test scan and wait for results strix cloud scans start --engagement-type live_test --domain-ids --wait # Example: list critical vulnerabilities as JSON strix cloud vulns list --severity critical --json ``` ## 🚀 Setup & Installation Strix is a Python/uv project. Install from the upstream repository: ```bash # Clone Strix (upstream: https://github.com/usestrix/strix) git clone https://github.com/usestrix/strix.git cd strix # Install dependencies uv sync # Run Strix uv run strix --help ``` Configure MCP servers (optional) in `~/.strix/mcp-servers.json`: ```json [ { "name": "github", "transport": "http", "url": "https://api.githubcopilot.com/mcp/", "auth": { "kind": "bearer", "token": "your-token" }, "allowed_tools": ["list_issues"] } ] ``` ## 📊 Result Handling - Strix cloud JSON output is machine-readable; use `--json` when stdout is not a terminal. - Binary downloads redirect raw bytes or use `--output FILE --json` for download metadata. ## 🛡️ Security & Red Lines - **Authorized use only:** Run Strix exclusively against systems you own or have explicit, written permission to test, within the agreed scope. - **You are responsible** for obtaining authorization and complying with applicable law. - **No third-party targets:** Never point Strix at applications, websites, or APIs you do not own or are not explicitly authorized to assess. - Provided "as is" with no warranty; misuse is your own responsibility. ## See Also - Strix repository: https://github.com/usestrix/strix - Strix documentation: https://docs.strix.ai - REA_Skill: https://github.com/smartworldarafath/REA_Skill