Skip to content
Back to search

claude-pentest-skills

Structured web application penetration testing with OWASP methodology, curated payload references, 6-gate validation, and professional report generation

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/frendysanusi--claude-pentest-skills/badge.svg)](/listings/frendysanusi--claude-pentest-skills)

HTML

<a href="/listings/frendysanusi--claude-pentest-skills"><img src="/api/listings/frendysanusi--claude-pentest-skills/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

**Step 1 — Clone**

git clone https://github.com/frendysanusi/claude-pentest-skills.git
cd claude-pentest-skills

**Step 2 — (Optional) Set up tools**

# Burp Suite MCP
cp .mcp.json.example .mcp.json
# Edit with your local paths

# PDF report generation
python3 -m venv .venv
.venv/bin/pip install markdown weasyprint

**Step 3 — Hunt**

claude                          # Start Claude Code in the project dir

/scope target.com               # Define authorization (REQUIRED first step)
/recon target.com               # Map attack surface
/hunt injection                 # Test for SQLi, NoSQLi, CMDi, SSTI, XXE
/validate                       # 6-gate quality check
/report --full                  # Generate pentest report

<br>

---

<br>

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

SKILL.md · Manifest · 108 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. ---
  2. name: claude-pentest-skills
  3. description: Structured web application penetration testing with OWASP methodology, curated payload references, 6-gate validation, and professional report generation
  4. ---
  5. # Claude Pentest Skills
  6. Structured methodology, curated payload references, and strict validation processes for **authorized** web application penetration testing.
  7. ## When to use
  8. Use this skill when performing authorized web application penetration testing. It provides:
  9. - A structured engagement workflow: scope definition, reconnaissance, vulnerability hunting, finding validation, and report generation
  10. - OWASP Web Security Testing Guide (WSTG) methodology with coverage tracking
  11. - 18 vulnerability class references with PayloadAllTheThings links
  12. - A 6-gate validation process that eliminates false positives
  13. - Professional pentest report generation (markdown + PDF)
  14. **Only use against systems you own or have explicit written authorization to test.**
  15. ## Instructions
  16. ### Standard Workflow
  17. ```
  18. /scope <target> → Define authorization (REQUIRED first step)
  19. /recon <target> → Map attack surface (endpoints, tech stack, inputs)
  20. /hunt <vuln-class> → Test for a specific vulnerability class
  21. /validate → Run 6-gate quality check on findings
  22. /report --full → Generate professional pentest report
  23. /status → View engagement progress dashboard
  24. ```
  25. ### Safety Rules (Non-Negotiable)
  26. 1. **Never** test without running `/scope` first
  27. 2. **Never** send requests to out-of-scope domains or paths
  28. 3. **Always** log every payload attempt before analyzing the response
  29. 4. **Stop and ask** the user if uncertain whether an action is in scope
  30. 5. **Warn and get confirmation** before testing destructive payloads (DELETE, data modification)
  31. 6. **Never** send attack payloads to third-party services (Auth0, AWS, etc.) — only test the client's own infrastructure
  32. ### Hunt Targets
  33. Valid arguments for `/hunt`:
  34. | Group | Includes |
  35. |-------|----------|
  36. | `injection` | SQLi, NoSQLi, CMDi, SSTI, XXE |
  37. | `auth-access` | IDOR, JWT, OAuth, privilege escalation |
  38. | `server-side` | SSRF, path traversal, file upload, deserialization, request smuggling |
  39. | `client-side` | XSS, CSRF, open redirect |
  40. | `session` | Session fixation, token weakness, cookie security |
  41. | `business-logic` | Race conditions, workflow bypass, logic flaws |
  42. | `config` | Security headers, CORS, TLS, info disclosure |
  43. | `ai-llm` | Prompt injection, LLM tool abuse, system prompt leakage |
  44. Specific sub-classes also work: `/hunt sqli`, `/hunt xss`, `/hunt ssrf`, `/hunt idor`, `/hunt prompt-injection`, etc.
  45. ### Agent Personas
  46. Three specialized personas are activated by commands:
  47. - **Recon Agent** (`/recon`) — Systematic mapper. Passive-first, methodical, records everything.
  48. - **Hunter Agent** (`/hunt`) — Creative attacker. Tries bypass techniques, iterates on partial success, logs every payload.
  49. - **Validator Agent** (`/validate`) — Skeptical critic. Assumes false positive until proven otherwise. Runs the 6-gate quality check.
  50. When no command is active, behave as a general security consultant.
  51. ### 6-Gate Validation
  52. Every finding must pass all 6 gates to be included in reports:
  53. 1. **Reproducible PoC** — Complete HTTP request that triggers the vulnerability
  54. 2. **HTTP Evidence** — Both request AND response captured
  55. 3. **Impact Verified** — Concrete and demonstrated, not theoretical
  56. 4. **In-Scope** — Endpoint is within authorized scope
  57. 5. **Real Vulnerability** — Not informational or best-practice observation
  58. 6. **Client Reproducible** — A third party can follow the documented steps
  59. ### Engagement State
  60. All engagement data is stored in `engagements/<target>/`:
  61. - `scope.md` — Authorization and scope boundaries
  62. - `findings.md` — Finding table with detailed evidence
  63. - `payload-log.md` — Append-only payload testing history
  64. - `coverage.md` — OWASP WSTG test coverage tracker
  65. - `tech-stack.md` — Discovered technologies
  66. - `notes.md` — Observations and hypotheses
  67. ### Tool Integration
  68. - **Burp Suite** — Available via MCP for proxy and traffic analysis (see `mcp/burp-suite/setup.md`)
  69. - **SQLMap** — Use via bash **only after** manual detection confirms SQL injection
  70. - **Commix** — Use via bash **only after** manual detection confirms command injection
  71. - **Custom scripts** — Generate and execute Python/Bash scripts for race conditions, token analysis, and custom payload delivery
  72. ### File Loading
  73. When a slash command is invoked, load files in this order:
  74. 1. `commands/<command>.md` — Command orchestration
  75. 2. `agents/<agent>.md` — Agent persona (if command activates one)
  76. 3. `skills/<skill>.md` — Methodology
  77. 4. `payloads/<payload>.md` — Payload reference (if hunting)
  78. 5. `engagements/<target>/<state-files>` — Current engagement state
  79. Load on demand, not all at startup.
README.md · README · 391 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. <div align="center">
  2. # Claude Pentest Skills
  3. ### Structured Web Application Penetration Testing with Claude Code
  4. *OWASP-based methodology, curated payload references, strict validation gates, and report generation — all orchestrated through slash commands.*
  5. <br>
  6. [![Claude Code](https://img.shields.io/badge/Claude_Code-Skill_Pack-D97706.svg?style=flat-square&logo=anthropic&logoColor=white)](https://docs.anthropic.com/en/docs/claude-code)
  7. [![OWASP WSTG](https://img.shields.io/badge/OWASP-WSTG_v4.2-0D47A1.svg?style=flat-square)](https://owasp.org/www-project-web-security-testing-guide/)
  8. <br>
  9. <a href="#-quick-start">Quick Start</a>&nbsp;&nbsp;|&nbsp;&nbsp;<a href="#-how-it-works">How It Works</a>&nbsp;&nbsp;|&nbsp;&nbsp;<a href="#-commands">Commands</a>&nbsp;&nbsp;|&nbsp;&nbsp;<a href="#-vulnerability-coverage">Coverage</a>&nbsp;&nbsp;|&nbsp;&nbsp;<a href="#-project-structure">Structure</a>&nbsp;&nbsp;|&nbsp;&nbsp;<a href="#-contributing">Contributing</a>
  10. <br>
  11. ```
  12. 6 slash commands · 3 agent personas · 18 vuln classes
  13. 6-gate validation · OWASP WSTG mapped · PDF report gen
  14. ```
  15. </div>
  16. <br>
  17. ---
  18. <br>
  19. ## The Problem
  20. Pentesting with AI is powerful, but without structure it's chaotic:
  21. - Claude sends payloads to **out-of-scope** targets
  22. - Findings are **false positives** that waste report-writing time
  23. - No consistent **methodology** — different approach every session
  24. - **Payload knowledge** scattered across bookmarks and notes
  25. - Reports are **written from scratch** every engagement
  26. <br>
  27. ## The Solution
  28. Claude Pentest Skills is a **structured skill pack** for Claude Code. It enforces scope, follows OWASP methodology, validates findings through a 6-gate quality process, and generates professional reports — all through simple slash commands.
  29. <br>
  30. <div align="center">
  31. | Before | After |
  32. |:---|:---|
  33. | Ad-hoc testing, no structure | OWASP WSTG methodology with coverage tracking |
  34. | False positives in reports | 6-gate validation kills bad findings before they waste your time |
  35. | "Is this in scope?" uncertainty | Mandatory `/scope` gate — refuses to test without authorization |
  36. | Write reports from scratch | `/report --full` generates professional markdown + PDF |
  37. | Forget what you tested | Payload log + coverage tracker across sessions |
  38. | Search PayloadAllTheThings manually | Curated methodology + direct references per vuln class |
  39. </div>
  40. <br>
  41. ---
  42. <br>
  43. ## Quick Start
  44. **Step 1 — Clone**
  45. ```bash
  46. git clone https://github.com/frendysanusi/claude-pentest-skills.git
  47. cd claude-pentest-skills
  48. ```
  49. **Step 2 — (Optional) Set up tools**
  50. ```bash
  51. # Burp Suite MCP
  52. cp .mcp.json.example .mcp.json
  53. # Edit with your local paths
  54. # PDF report generation
  55. python3 -m venv .venv
  56. .venv/bin/pip install markdown weasyprint
  57. ```
  58. **Step 3 — Hunt**
  59. ```bash
  60. claude # Start Claude Code in the project dir
  61. /scope target.com # Define authorization (REQUIRED first step)
  62. /recon target.com # Map attack surface
  63. /hunt injection # Test for SQLi, NoSQLi, CMDi, SSTI, XXE
  64. /validate # 6-gate quality check
  65. /report --full # Generate pentest report
  66. ```
  67. <br>
  68. ---
  69. <br>
  70. ## How It Works
  71. ```
  72. YOU
  73. |
  74. ┌─────▼─────┐
  75. │ Claude │ <── Burp MCP (optional)
  76. │ Code │
  77. └─────┬─────┘
  78. |
  79. ┌───────────────┼───────────────┐
  80. | | |
  81. ┌─────▼──────┐ ┌──────▼─────┐ ┌─────▼─────┐
  82. │ Recon │ │ Hunter │ │ Validator │
  83. │ Agent │ │ Agent │ │ Agent │
  84. └─────┬──────┘ └──────┬─────┘ └─────┬─────┘
  85. | | |
  86. curl / Burp payloads/ 6-gate check
  87. JS analysis PASS / REJECT
  88. tech fingerprint NEEDS-WORK
  89. | | |
  90. ┌─────▼───────────────▼───────────────▼──────┐
  91. │ engagements/<target>/ │
  92. │ │
  93. │ scope.md · findings.md · payload-log.md │
  94. │ coverage.md · tech-stack.md · notes.md │
  95. └────────────────────────────────────────────┘
  96. |
  97. ┌─────▼─────┐
  98. │ /report │ ──> Markdown + PDF
  99. └───────────┘
  100. ```
  101. Each command loads the right agent persona, methodology skill, and payload references. State is tracked per engagement in `engagements/<target>/`.
  102. <br>
  103. ---
  104. <br>
  105. ## Commands
  106. | Command | What It Does |
  107. |:---|:---|
  108. | `/scope <target>` | Define target scope and confirm authorization. **Must run first.** |
  109. | `/recon <target>` | Map attack surface — tech stack, endpoints, input vectors, WAF detection |
  110. | `/hunt <vuln-class>` | Test for a specific vulnerability class using OWASP methodology |
  111. | `/validate` | Run 6-gate quality check on all unvalidated findings |
  112. | `/report --full` | Generate complete pentest report (exec summary, findings, remediation) |
  113. | `/report --finding F-001` | Generate report for a single finding |
  114. | `/status` | Show engagement dashboard — findings, coverage, payload stats |
  115. <br>
  116. ### Hunt Targets
  117. ```bash
  118. /hunt injection # SQLi, NoSQLi, CMDi, SSTI, XXE
  119. /hunt auth-access # IDOR, JWT, OAuth, privilege escalation
  120. /hunt server-side # SSRF, path traversal, file upload, deserialization
  121. /hunt client-side # XSS, CSRF, open redirect
  122. /hunt session # Session fixation, token weakness, cookie security
  123. /hunt business-logic # Race conditions, workflow bypass, logic flaws
  124. /hunt config # Headers, CORS, TLS, info disclosure
  125. /hunt ai-llm # Prompt injection, LLM tool abuse, system prompt leakage
  126. # Or target specific sub-classes:
  127. /hunt sqli /hunt xss /hunt ssrf
  128. /hunt idor /hunt jwt /hunt prompt-injection
  129. ```
  130. <br>
  131. ---
  132. <br>
  133. ## Agent Personas
  134. 3 specialized personas, each activated by the relevant command:
  135. | Agent | Activated By | Personality | Role |
  136. |:---|:---|:---|:---|
  137. | **Recon Agent** | `/recon` | Systematic, methodical | Maps the full attack surface before exploitation |
  138. | **Hunter Agent** | `/hunt` | Creative, persistent | Tries bypass techniques, iterates on partial success |
  139. | **Validator Agent** | `/validate` | Skeptical, adversarial | Assumes false positive until proven otherwise |
  140. When no command is active, Claude behaves as a general security consultant.
  141. <br>
  142. ---
  143. <br>
  144. ## Validation Gate
  145. Every finding must pass **all 6 gates** to reach the report. One failure = REJECTED or NEEDS-WORK.
  146. ```
  147. Finding ──> Gate 1: Reproducible PoC?
  148. Gate 2: HTTP evidence captured?
  149. Gate 3: Impact verified (not theoretical)?
  150. Gate 4: Target in scope?
  151. Gate 5: Real vulnerability (not informational)?
  152. Gate 6: Client can reproduce it?
  153. │
  154. ┌──────────┼──────────┐
  155. │ │ │
  156. VALIDATED REJECTED NEEDS-WORK
  157. (→ report) (killed) (fix & retry)
  158. ```
  159. This process eliminates false positives, ensures evidence quality, and produces findings that clients can verify independently.
  160. <br>
  161. ---
  162. <br>
  163. ## Vulnerability Coverage
  164. <details>
  165. <summary><b>18 Vulnerability Classes</b> — click to expand</summary>
  166. <br>
  167. | Class | Sub-Classes |
  168. |:---|:---|
  169. | **SQL Injection** | Union, blind, time-based, error-based, auth bypass |
  170. | **XSS** | Reflected, stored, DOM, CSP bypass, polyglots |
  171. | **SSRF** | Internal access, cloud metadata, protocol abuse, IP bypass |
  172. | **SSTI** | Jinja2, Twig, Freemarker, Pebble, Mako, ERB |
  173. | **Command Injection** | Metacharacters, blind detection, space/quote bypass |
  174. | **XXE** | Classic, blind OOB, SVG upload, XInclude, PHP filter |
  175. | **Path Traversal** | Encoding bypass, null byte, overlong UTF-8 |
  176. | **File Upload** | Extension bypass, MIME spoof, magic bytes, polyglots |
  177. | **IDOR** | Sequential ID, UUID harvest, GraphQL node, HPP |
  178. | **JWT Attacks** | alg:none, RS256→HS256, KID injection, claim tampering |
  179. | **OAuth/OIDC** | Redirect URI bypass, state absence, PKCE downgrade |
  180. | **NoSQL Injection** | MongoDB operator injection, $regex blind, $where |
  181. | **Deserialization** | Java/PHP/Python/.NET/Ruby/YAML gadgets |
  182. | **CSRF** | Auto-submit forms, JSON CSRF, SameSite bypass |
  183. | **Open Redirect** | Protocol-relative, encoding chains, @ symbol tricks |
  184. | **Request Smuggling** | CL.TE, TE.CL, HTTP/2 downgrade, TE obfuscation |
  185. | **Race Condition** | TOCTOU, coupon reuse, rate limit bypass, double spend |
  186. | **Prompt Injection** | Direct override, system prompt extraction, persona hijack, encoding bypass, tool abuse, indirect injection |
  187. </details>
  188. <br>
  189. ---
  190. <br>
  191. ## Project Structure
  192. ```
  193. claude-pentest-skills/
  194. ├── commands/
  195. │ ├── scope.md # /scope — define target authorization
  196. │ ├── recon.md # /recon — map attack surface
  197. │ ├── hunt.md # /hunt — test vulnerability class
  198. │ ├── validate.md # /validate — 6-gate quality check
  199. │ ├── report.md # /report — generate pentest report
  200. │ └── status.md # /status — engagement dashboard
  201. │
  202. ├── agents/ # Agent persona definitions
  203. │ ├── recon-agent.md
  204. │ ├── hunter-agent.md
  205. │ └── validator-agent.md
  206. │
  207. ├── skills/
  208. │ ├── recon.md # Reconnaissance phases
  209. │ ├── validate.md # 6-gate validation process
  210. │ ├── report.md # Report templates + CVSS guide
  211. │ └── hunting/ # Per-vuln-class methodology
  212. │ ├── injection.md │ auth-access.md │ server-side.md
  213. │ ├── client-side.md │ session.md │ business-logic.md
  214. │ ├── config.md │ ai-llm.md
  215. │
  216. ├── payloads/ # Methodology + payload references
  217. │ ├── sqli.md │ xss.md │ ssrf.md │ ssti.md │ xxe.md
  218. │ ├── cmdi.md │ idor.md │ jwt.md │ oauth.md │ nosqli.md
  219. │ ├── csrf.md │ path-traversal.md │ file-upload.md
  220. │ ├── deserialization.md │ open-redirect.md
  221. │ ├── request-smuggling.md│ race-condition.md
  222. │ └── prompt-injection.md
  223. │
  224. ├── mcp/ # MCP server integration docs
  225. │ └── burp-suite/ # Burp Suite proxy setup
  226. │
  227. ├── tools/ # Utilities
  228. │ └── md2pdf.py # Markdown → styled PDF report
  229. │
  230. ├── engagements/ # Per-target state (gitignored)
  231. │ └── <target>/
  232. │ ├── scope.md │ findings.md │ payload-log.md
  233. │ ├── coverage.md │ tech-stack.md │ notes.md
  234. │ └── report-*.md / report-*.pdf
  235. │
  236. ├── local-payloads/ # Full curated payloads (gitignored)
  237. ├── CLAUDE.md # Master config for Claude Code
  238. └── .mcp.json.example # MCP config template
  239. ```
  240. <br>
  241. ---
  242. <br>
  243. ## Tool Integration
  244. | Tool | Integration | Use |
  245. |:---|:---|:---|
  246. | **Burp Suite** | MCP server | Proxy, traffic analysis, request replay |
  247. | **SQLMap** | Bash (post-detection) | Automated SQL injection exploitation |
  248. | **Commix** | Bash (post-detection) | Automated command injection exploitation |
  249. | **Custom scripts** | Claude generates on the fly | Race conditions, token analysis, encoding |
  250. | **md2pdf.py** | Python (weasyprint) | Convert report markdown to styled PDF |
  251. > SQLMap and Commix are used **only after manual detection** confirms the vulnerability. Never as first-pass scanners.
  252. <br>
  253. ---
  254. <br>
  255. ## Local Payloads
  256. The public `payloads/` directory contains methodology and payload references. For curated payload lists with full exploit code, add them to `local-payloads/` (gitignored):
  257. ```bash
  258. mkdir local-payloads
  259. # Add your curated .md files with full payloads here
  260. # The hunt skills will reference both payloads/ and local-payloads/
  261. ```
  262. <br>
  263. ---
  264. <br>
  265. ## Safety Rules
  266. These are enforced in `CLAUDE.md`. Non-negotiable.
  267. ```
  268. 1. SCOPE FIRST /scope must run before any other command
  269. 2. CHECK EVERY REQUEST verify against scope.md before sending
  270. 3. LOG BEFORE ANALYZE write payload to log BEFORE reading response
  271. 4. NO OUT-OF-SCOPE refuse requests to excluded domains/paths
  272. 5. ASK IF UNCERTAIN stop and confirm with user when scope is ambiguous
  273. 6. WARN ON DESTRUCTIVE get explicit approval for DELETE/data modification
  274. ```
  275. <br>
  276. ---
  277. <br>
  278. ## Contributing
  279. Contributions are welcome. Please read **[CONTRIBUTING.md](CONTRIBUTING.md)** before opening a pull request.
  280. <br>
  281. ---
  282. <br>
  283. <div align="center">
  284. **For authorized security testing only.**
  285. <br>
  286. Only test systems you own or have explicit written permission to test.
  287. </div>

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.