What problem does it solve? Manual bug bounty hunting requires repeatedly coordinating scope verification, reconnaissance, endpoint ranking, vulnerability testing, and report drafting, which is slow and error-prone across multiple targets. ## Core Features & Use Cases - Autonomous Hunt Loop: Chains scope check, recon, attack surface ranking, hunting, 7-question validation, and report drafting into one continuous workflow. - Configurable Checkpoint Modes: Choose --paranoid, --normal, --yolo, or --quick to control how often the loop pauses for human review and how many tokens it consumes. - Safety Guarantees: Every URL is checked against a scope allowlist, all requests are logged to hunt-memory/audit.jsonl, rate limits and a circuit breaker prevent hammering, and reports are never auto-submitted. - Use Case: A bug bounty hunter points the skill at target.com in --paranoid mode on a new program, letting it recon and test P1 endpoints while pausing on every finding for human approval. ## Quick Start Ask the agent to run /autopilot target.com in a fresh session to start a full scoped hunt with paranoid checkpoints.
