@ellistevo/openclaw-pentest
v1.0.0
Published
AI-driven penetration testing CLI - point it at your app and it tries to break in
Downloads
18
Maintainers
Readme
@ellistevo/openclaw-pentest
AI-driven penetration testing CLI. Point it at your app and it tries to break in.
⚠️ Disclaimer
Only test applications you own or have explicit written permission to test. Unauthorized testing is illegal.
Install & Run
npx @ellistevo/openclaw-pentest https://myapp.comOr install globally:
npm install -g @ellistevo/openclaw-pentest
openclaw-pentest https://myapp.comTests Performed
| Test | What it checks | |------|---------------| | info-disclosure | Server headers, stack traces in error pages | | exposed-files | .env, .git, backups, phpinfo, actuator, swagger, etc. | | security-headers | HSTS, CSP, X-Frame-Options, etc. | | cors | Wildcard/reflected origins, credential leaks | | xss | Reflected XSS via common parameters | | sqli | SQL injection error-based detection | | open-redirect | Unvalidated redirects via URL parameters | | directory-traversal | LFI/path traversal attacks | | auth-bypass | Unauthenticated admin access, JWT detection | | rate-limiting | Missing rate limits on general + login endpoints | | prompt-injection | AI endpoint detection + injection attempts | | csrf | Missing CSRF tokens, SameSite cookies |
Options
-o, --output <dir> Output directory (default: ./pentest-report)
-t, --timeout <ms> Request timeout (default: 10000)
--skip <tests> Comma-separated tests to skip
--only <tests> Comma-separated tests to run
-v, --verbose Verbose output
--json-only Only generate JSON reportOutput
Reports are saved to ./pentest-report/:
report.html— Beautiful dark-themed HTML reportreport.json— Machine-readable JSON
Safety
All tests are non-destructive. No data is modified, no accounts are created, no files are uploaded. The tool only sends HTTP requests and analyzes responses.
