Agent skills for Claude Code. Start at AGENTS.md.
/recon example.com # run recon against a target
/recon example.com --fresh # re-run all phases from scratch
/recon-setup # install missing tools and wordlists
AGENTS.md ← start here (table of contents)
recon/SKILL.md ← architecture, phase graph, output contracts
recon/references/<phase>.md ← one self-contained guide per phase
recon/scripts/ ← helper scripts
.claude/commands/recon.md ← orchestrator entrypoint
results/<target>/ ← output (gitignored)
The orchestrator observes existing state, builds a plan, confirms with the user, then dispatches sub-agents. See AGENTS.md for principles.