OSINT Skill by smixs is an open-source intelligence-gathering skill for AI coding agents (Claude Code, OpenClaw, and compatible SKILL.md platforms). Given minimal input like a name or handle, it runs a multi-phase investigation — from quick web searches through deep social media extraction — using 55+ Apify scraping actors and seven search APIs, and outputs a scored dossier with a psychoprofile, career map, and confidence grades per finding. It's built for legitimate research and due-diligence use cases; treat it the same as any OSINT tool with respect to consent and applicable law.
OSINT Skill by smixs is an open-source intelligence-gathering skill for AI coding agents, packaged in the SKILL.md format that Claude Code, OpenClaw and compatible platforms load. You install it into your agent, give it minimal input such as a name or a handle, and it runs a staged investigation: quick web searches first, then deeper extraction from social platforms as the picture fills in. It drives that work through a large set of Apify scraping actors and several search APIs, meaning the collection is done by external services the skill orchestrates rather than by code it runs locally. The output is a structured dossier rather than a pile of links, with a psychoprofile section, a career map and a confidence grade attached to each finding so you can see what is solid and what is inference. It is built for legitimate research and due diligence, and it carries the obligations any OSINT tooling does around consent, data protection law and how results are used.
This fits investigative work you would otherwise do by hand across a dozen tabs: vetting a business counterparty, checking a potential hire or partner against public records, journalism background work, or verifying that a person presenting themselves online is who they claim. The per-finding confidence grading is the part that earns its keep, because it forces a distinction between confirmed facts and plausible matches, which is where manual OSINT most often goes wrong. It assumes you already run an agent and can supply the API credentials it depends on. Do not treat it as a general people-search toy. Profiling individuals carries legal and ethical weight, and in many jurisdictions the purpose you collect for is itself regulated, so confirm you have a lawful basis before running it.
Anthropic Cybersecurity Skills is the nearest neighbor in format, a set of agent skills for security work, but its focus is defensive and technical rather than building a profile of a person, so the two address different halves of an investigation. User Research Skill also produces structured findings about people, though it is aimed at understanding users and their needs for product decisions, using consented research rather than open-source collection about a named individual. Crawl4AI is the raw ingredient rather than the method: a crawler that turns web pages into clean, model-ready text, which you would use if you were assembling your own pipeline. OSINT Skill's difference is the prescribed multi-phase method and the graded dossier at the end.
Compare with: Anthropic Cybersecurity Skills, User Research Skill, Crawl4AI
The catalog records OSINT Skill as free, and the skill itself is open source with no purchase involved. Running it is not free, though. It depends on third-party scraping actors and search APIs, and those services bill on usage, so a deep investigation with many phases means many paid calls on accounts you supply. Model tokens are an additional cost, and agentic multi-step work consumes them quickly. Before a first run, set spending limits on the underlying service accounts, because an investigation that expands on its own findings can make far more requests than you expected.
No notes have been added for this tool yet.
No prompting tips have been added yet.
No reviews yet. Be the first!