autopilot · h-mmer — Claude Code skill: autonomous hunt orchestrator
autopilot — a Claude Code skill that runs a hunt across all vulnerability classes for a target and keeps the agent from calling it done too early.
no license ★ 1,298 updated 2026-06-12
npx skills add H-mmer/pentest-agents --skill autopilot -a claude-codeWhat autopilot does
Point autopilot at a host and it assigns each vulnerability class its own hunter subagent: IDOR, the XSS variants, SSRF, SQLi and the rest of the fixed class list. A class may only be skipped with a technical reason tied to what the surface probe found, recorded as a not-applicable entry in brain state.
Setup is strict. The skill loads rules/hunting.md and rules/mistakes.md, confirms targets against scope.yaml, then pulls every constraint out of policy.md and injects that block into each subagent dispatch — headers, account rules, forbidden actions. Run metadata lands in .autopilot-run.env, prior knowledge comes from uv run python3 $CLAUDE_PROJECT_DIR/tools/brain.py brief <target>.
--autonomous enforces an exhaustion contract: the loop keeps dispatching until every target is tested and the queues are empty, with no clarifying questions. That makes a complete policy.md and an accurate scope.yaml the difference between a real run and a gate failure.
What it helps with
- cover every vulnerability class on a target, not just the obvious ones
- record a written reason when a class does not apply
- carry program rules into each subagent dispatch
- pick up an interrupted hunt from saved brain state
- produce a submission-ready report without auto-submitting
Example prompts for the agent
/autopilot target.com --interactive— run the hunt and pause after each validated finding/autopilot target.com --autonomous— keep going until every class is exhausted/autopilot target.com --resume— continue from the previous session
How to install autopilot
1Skills CLI (Vercel)
npx skills add H-mmer/pentest-agents --skill autopilot -a claude-code2Manually: put the files in place
~/.claude/skills/autopilot/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/autopilot/
1Skills CLI (Vercel)
npx skills add H-mmer/pentest-agents --skill autopilot -a codex2Manually: put the files in place
.agents/skills/autopilot/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add H-mmer/pentest-agents --skill autopilot -a cursor2Manually: put the files in place
.cursor/skills/autopilot/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install H-mmer/pentest-agents .claude/skills/autopilot2Manually: put the files in place
.github/skills/autopilot/SKILL.md1Skills CLI (Vercel)
npx skills add H-mmer/pentest-agents --skill autopilot -a gemini-cli2Platform CLI
gemini skills install https://github.com/H-mmer/pentest-agents --path .claude/skills/autopilotManually: put the files in place
.devin/skills/autopilot/SKILL.mdThe legacy .windsurf/skills/ path still works
Claude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about autopilot
How do I install autopilot in Claude Code?
Run npx skills add H-mmer/pentest-agents --skill autopilot -a claude-code. The skill file is placed at ~/.claude/skills/autopilot/SKILL.md.
What does autopilot do?
It orchestrates testing of a target across a fixed set of vulnerability classes, dispatching hunter subagents and tracking exhaustion before a summary.
What is the difference between --interactive and --autonomous?
--interactive pauses after each validated finding so you can review it. --autonomous runs with no pauses, though it never submits reports automatically.
Similar skills
All skills for Security →autopilot by other authors comes first.