skills.rest
greglas75
greglas75COMMUNITY
Explore all skills by this author →
GitHub

pentest

Runs hybrid white-box and black-box penetration tests across seven security dimensions with source-to-sink tracing.

6|Updated Mar 27, 2026
Search Tags:#security#vulnerability#penetration testing#black-box#white-box#source-to-sink#exploit verification
One-click install
npx skills add https://github.com/greglas75/zuvo --skill pentest-greglas75
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: pentest
Source: https://github.com/greglas75/zuvo/tree/main/skills/pentest
Command: npx skills add https://github.com/greglas75/zuvo --skill pentest-greglas75

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Security reviews often stop at static warnings without proving exploitability, or rely on black-box scans that miss code-level context. This Skill combines white-box source-to-sink tracing with optional runtime exploit verification, producing deduplicated, scored findings with file:line evidence. ## Core Features & Use Cases - Seven-dimension coverage (PT1-PT7): Injection, XSS, auth/session, authorization/IDOR, SSRF, input vectors, and business logic flaws, plus optional runtime checks and CMS overlays for WordPress, Joomla, and Drupal. - Stack-aware analysis: Detects frameworks (Django, Next.js, FastAPI, NestJS, Rails-style stacks) and selects matching sources, sinks, safe patterns, and CodeSift tools before dispatching agents. - Deterministic finding pipeline: Candidate schemas, canonical-key deduplication, confidence filters, MUST-GATE enforcement, coverage gates, and centralized scoring with grades and verdicts. - Use Case: Before a release, run a pentest against your staging URL and codebase to verify that flagged SQL injection and broken access control paths are actually exploitable, then receive a scored report with remediation evidence. ## Quick Start Ask the AI to run a penetration test on the current project with the staging URL, for example: run zuvo:pentest on this repo with --url http://localhost:3000 and verify the findings. Safety gates require explicit authorization, block production targets by default, cap request rates, and scrub PII from output.

Frequently Asked Questions about pentest

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run a penetration test on my codebase with this Skill?▼

Invoke zuvo:pentest with a path for white-box source analysis, or add --url <url> for hybrid mode with runtime probing. Use --quick for critical dimensions only, or --verify-live to confirm exploits against a running app.

What is the difference between white-box and black-box mode?▼

White-box traces attacker-controlled input to dangerous sinks in source code using call-chain analysis. Black-box dispatches scanner agents against a live URL through recon, experiment, test, and verify cycles. Hybrid mode combines both when source and URL are available.

Can it verify findings from a previous security audit?▼

Yes. The --from-audit <dir> flag imports PT/SEC findings from a prior audit, normalizes them into candidate records, deduplicates via cross-skill reconciliation, and re-verifies them with runtime checks where possible.

Does it support WordPress, Joomla, or Drupal sites?▼

Yes. CMS detection runs automatically via marker files like wp-config.php, or force it with --cms wordpress|joomla|drupal. The overlay checks version CVEs, extension vulnerabilities, admin exposure, config, and file permissions.

Is it safe to run against a production URL?▼

No by default. The skill refuses targets that are not localhost, staging, or dev unless explicitly overridden, requires authorization confirmation, caps request rates, uses read-only payloads, and scrubs PII from all output.

When should I use security-audit instead of pentest?▼

Use security-audit for broad security posture review and code-audit for static quality. Pentest is for proving exploitability of specific vulnerability classes through source-to-sink traces and optional live verification.

Related Skills

Explore More Software Engineering Skills→
UiPathUiPath
18

fix-security-vulnerabilities

Trace transitive dependencies and apply validated fixes for npm and pnpm vulnerabilities.

Official
Advanced
AlmogMaayanAlmogMaayan

ptp-receiving-code-review

Validates code review findings against current repository content before applying fixes.

Community
Intermediate
HomericIntelligenceHomericIntelligence
18

scan-vulnerabilities

Scan code and dependencies for security vulnerabilities using Bandit, Safety, and Semgrep.

Official
Intermediate
4lerman4lerman

security-pen-testing

Audit web applications, APIs, and infrastructure for security vulnerabilities.

Community
Advanced