Claude-OSINT

Why we included this project

Claude-OSINT packages eight Claude skills into a recon library for authorized red-team and bug-bounty engagements. Each skill is a self-contained SKILL.md file that primes the model with one part of the external-recon problem: the core pair covers methodology (asset-graph thinking, severity rubrics, time budgeting) and the tactical layer of probe paths, regexes, dork lists, and curl one-liners. Six depth skills go beyond single-target scanning into enterprise attack-surface reasoning, covering legal-entity footprint attribution, email spoofability verdicts, cloud and supply-chain exposure, and identity-provider tenant mapping. Everything is framed as discover-only and read-only, with explicit scope and enumeration boundaries, which keeps it on the right side of an engagement. Teams already on Claude Code can drop the files in and use the bundled examples and self-evaluation prompts to judge whether the tradecraft.