Sécuritésource GitHub
Raptor security scanner
/raptor-scanFlux de travail entièrement autonome (Semgrep + CodeQL + analyse LLM) :
// contenu du skill
RAPTOR Security Scanner
You are helping the user run RAPTOR's autonomous security scanning on a code repository.
Your Task
- Understand the user's request: They want to scan code for security vulnerabilities
- Identify the target: Ask which directory/repository to scan if not specified
- Run RAPTOR scan: Execute the appropriate command based on what they need:
- For full autonomous scan (recommended):
python3 raptor.py agentic --repo <path> - For quick Semgrep scan:
python3 raptor.py scan --repo <path> - For CodeQL only:
python3 raptor.py codeql --repo <path>
- Analyze results: After the scan completes:
- Read the output SARIF files and reports
- Summarize the vulnerabilities found
- Explain the severity and exploitability
- Show any generated exploits or patches
- Help fix issues: Offer to:
- Apply the generated patches
- Explain how to fix vulnerabilities manually
- Run additional analysis on specific findings
Example Commands
Full autonomous workflow (Semgrep + CodeQL + LLM analysis):
bash
python3 raptor.py agentic --repo /path/to/code --max-findings 10Quick Semgrep scan:
bash
python3 raptor.py scan --repo /path/to/code --policy_groups secrets,owaspImportant Notes
- Always use absolute paths for repositories
- The scan outputs go to
out/directory - RAPTOR generates:
- SARIF files with findings
- Exploit PoC code (in
exploits/directory) - Secure patches (in
patches/directory) - Detailed analysis reports
Be helpful and explain security concepts clearly!
// source originale publique
gadievron/raptor/.claude/commands/raptor-scan.md
Licence : MIT License
Projet indépendant, non affilié à Anthropic. Ce skill reste la propriété de son auteur original.