What problem does it solve? Manually gathering open-source intelligence on an entity (domain, IP, email, person, or company) means juggling dozens of disconnected tools like Whois, DNS, crt.sh, Wayback Machine, and Shodan, then stitching results together by hand. This Skill automates that pipeline into a single structured dossier with an audit trail, while enforcing strict safety rules: secrets are redacted, dump/leak sites are blocklisted, and raw responses are never written to disk unfiltered. ## Core Features & Use Cases - Multi-entity classification and recon: Automatically detects whether the target is a domain, IP, email, person, company, or GitHub user and runs the appropriate channels (Whois/DNS, crt.sh subdomains, Wayback snapshots, Shodan InternetDB, GitHub code search, optional theHarvester/subfinder). - Hybrid report generation: Produces a findings summary ranked CRITICAL to LOW, an entity dossier, a mermaid relationship graph, CSV artifacts (subdomains, emails, IPs), and a sources audit trail in .firecrawl/osint/<slug>/. - Built-in security guardrails: All channel output passes through an inbound blocklist filter and a secret redactor (AWS keys, GitHub tokens, Slack tokens, JWTs, private keys) before any disk write. - Use Case: Run /osint-research example.com before a vendor assessment to get a 10-15 minute passive recon report covering subdomains, exposed ports, leaked credentials in public code, and historical snapshots — with zero paid subscriptions. ## Quick Start Ask the AI to run an OSINT recon on a target, for example: run osint-research on example.com and give me the findings report.
