hunt
Test for a specific vulnerability class against target endpoints using curated payloads.
Map the attack surface of an authorized target.
$ npx -y skills add frendysanusi/claude-pentest-skills --agent claude-codeHow it fires
How this command gets triggered: by you, by Claude, or both.
/reconContext preview
What this command does when you run it.
Map the attack surface of an authorized target.
Map the attack surface of an authorized target.
/recon example.com /recon https://app.example.com
1. **Scope must exist.** Check for `engagements/<target>/scope.md`. If not found:
2. **Verify target matches scope.** The target must be listed as in-scope.
1. Load `agents/recon-agent.md` — Adopt the Recon Agent persona 2. Load `skills/recon.md` — Reconnaissance methodology
Read engagement state files:
If recon was partially completed before, resume from where it left off (check coverage.md).
Follow `skills/recon.md` phases 1 through 5:
1. **Technology Fingerprinting** → Update `tech-stack.md` 2. **Endpoint Discovery** → Update `coverage.md` endpoint table 3. **Input Vector Mapping** → Update `coverage.md` with input vectors per endpoint 4. **Authentication Mechanism ID** → Update `tech-stack.md` auth section 5. **Infrastructure & WAF Detection** → Update `tech-stack.md` CDN/WAF section
Generate the structured recon summary as defined in `skills/recon.md`:
Suggest the next command:
Recon complete for <target>. Discovered <X> endpoints, <Y> input vectors. Tech stack: <summary> Suggested next step: /hunt <recommended-vuln-class>
If the user runs `/recon` again on a target with existing recon data: 1. Show what phases are already complete 2. Ask: "Resume from Phase X, or start fresh?" 3. If resuming, skip completed phases 4. If starting fresh, clear and rebuild state files (confirm with user first)
AI-powered web application penetration testing skills with Claude Code
Repo: frendysanusi/claude-pentest-skills
Test for a specific vulnerability class against target endpoints using curated payloads.
Define and authorize the target scope for a penetration testing engagement. This is the **mandatory first step** — all other commands refuse to run without an…
Display the engagement progress dashboard. Read-only — does not modify any state files.