What problem does it solve? When a penetration testing engagement starts, an operator must decide which of dozens of attack skills to load for a given target. Loading everything exhausts the context window, while loading too little misses the platform-specific attack surface. This Skill automates that decision by fingerprinting live hosts and selecting the right skill set based on the engagement mode. ## Core Features & Use Cases - Target Fingerprinting: Runs curl-based header and body probes against every live host, following redirects, and maps signals (Okta, Entra, Next.js, Laravel, Kubernetes, gRPC, and more) to specific platform attack skills. - Mode-Based Skill Loading: Loads the always-on, platform, and hunt-* skill sets for redteam mode, or the full OWASP-oriented hunt-* set for WAPT blackbox/greybox mode, with an 8-skill load budget, tier-based priority, and de-duplication rules. - Greybox Credential Preflight: Validates supplied credentials with a low-cost authenticated probe before running authenticated tests, confirming the session is live, the role is known, and no MFA challenge blocks access. - Use Case: After answering the /hunt mode prompt with "wapt greybox", this Skill fingerprints the target, detects a Next.js front end behind an Okta SSO redirect, loads the matching skills within budget, validates the supplied session cookie, and prints the loaded-skill taxonomy before handing control back to /hunt. ## Quick Start After answering the /hunt mode prompt, invoke hunt-dispatch with your chosen mode such as mode=redteam or mode=wapt box=greybox to fingerprint the target and load the matching attack skills.