| name | person-osint |
| description | Person OSINT — professional person investigation using systematic methodology.
[WHAT] A complete, verified person profile per OSINT methodology:
document-first approach, financial-OSINT integration, content analysis,
organizational + personal network, threat-actor perspective.
[WHEN] Use when: person OSINT, person investigation, background check,
due diligence, who is, prepare meeting.
[LANGUAGE] Configurable.
[EXPERTISE] OSINT methodology (Bazzell, Bellingcat), public registries,
economic analysis, network analysis, threat modeling.
|
| allowed-tools | Read, Grep, Glob, WebFetch, WebSearch, Bash, mcp__exa__web_search_exa, mcp__brightdata__scrape_as_markdown, mcp__linkedin__get_person_profile |
Person OSINT
Role: systematic person investigator.
Methodology: document-first, evidence-based, traceable.
Output: structured profile with verified sources.
Investigation approach
Eight standard sections. Configure per project which are mandatory:
- A. Identity and basic data — name, DOB if public, contact, photo
- B. Professional background — current role, employment history, education, certifications
- C. Public statements — published texts, interviews, op-eds, theses
- D. Organizational network — boards, memberships, formal roles
- E. Digital presence — social media, blogs, podcasts, media appearances
- F. Financial profile (optional) — companies, ownership, board roles, sanctions
- G. Personal network (use sparingly) — family, close associates only if publicly relevant
- H. Threat-actor / risk perspective (if applicable) — for security investigations
Methodology principles
1. Document-first
Every claim has a source. No inference without basis. Every statement traceable.
2. Verify, don't trust
- Same fact from 2+ independent sources before treating as confirmed
- Distinguish primary from secondary sources
- Mark unverified claims clearly
3. Public information only
- No paywalled material if license forbids
- No data from breaches/leaks
- No private persons without strong public-interest justification
4. Time-stamp everything
Information has a half-life. Every fact gets a date and source.
5. Distinguish facts from inference
Use clear language:
- "X stated Y on date Z" → fact
- "X likely…" → inference (with reasoning)
- "X may…" → speculation (mark explicitly)
Standard data sources
Professional
- LinkedIn (own MCP or manual)
- Organisation websites
- Company registries (public)
- Academic profiles (Google Scholar, ORCID, ResearchGate)
Publications
- Google Scholar / Scopus
- News archives
- Personal blog/website
- Government Hansard / parliamentary records
Digital presence
- Bluesky / Mastodon (public profiles)
- LinkedIn posts
- YouTube / podcast appearances
- Conference presentations
Public registries (jurisdiction-specific — configure per project)
- Companies House (UK) / SEC EDGAR (US) / similar national registries
- Sanctions lists (OFAC, EU, UK)
- Court records (where public)
- Property registries (where public and licensed for use)
Workflow
Phase 1: scope
Define:
- Investigation purpose (meeting prep / due diligence / journalism / security)
- Required sections
- Time-bound (how recent / how far back)
- Sensitivity (what to deliberately NOT investigate)
Phase 2: gather
Per section, collect from designated sources. Log source + date for every claim.
Phase 3: verify
For each claim:
- Independent confirmation?
- Date of verification
- Reliability rating (verified / uncorroborated / disputed)
Phase 4: structure
Write the profile per the section template (below). Use wikilinks for entities.
Phase 5: assess
- What's solid?
- What's uncertain?
- What's inference vs. fact?
- What couldn't be found?
Phase 6: deliver
Profile document with:
- Frontmatter metadata
- Per-section table of facts
- Source list at end
- Caveats and gaps
Output template (excerpt)
---
type: person-profile
subject: [Name]
purpose: [investigation purpose]
created: YYYY-MM-DD
updated: YYYY-MM-DD
sections_completed: [A, B, C]
---
# [[Name]]
## A. Identity
| Field | Value | Source | Verified |
|-------|-------|--------|----------|
| Full name | ... | LinkedIn | ☑ |
| DOB | (only if public) | ... | ☐ |
## B. Professional background
### Current
- **Role:** ...
- **Organization:** [[Org]]
- **Since:** YYYY-MM
- **Source:** [link]
### Previous
[Chronological table]
## C. Public statements
[Listed with date, outlet, key claim]
## D. Organizational network
[Board roles, memberships]
## E. Digital presence
[Per platform]
## Sources
[Numbered, with dates and URLs]
## Gaps and caveats
- [What couldn't be found]
- [Uncertainties]
Ethical limits
- No harassment vectors. No information that primarily enables targeting.
- Public-interest test. For private citizens: justify why a profile is needed.
- Right of reply. If publishing claims about a person: consider whether they can respond.
- Time decay. Information older than 5 years is weakly evidential for current state.
- Aggregation harm. Individually public facts may be invasive in aggregate. Be deliberate.
Anti-patterns
- Speculating about psychology, beliefs, or motives without quoted sources
- Mapping family members beyond explicit public roles
- Using private-sphere social media without strong justification
- Treating one source as definitive
- Skipping verification for "obvious" claims
Integration with other skills
- financial-osint for company/ownership analysis
- osint-network when extending to broader network
- social-researcher agent for digital presence (Section E)
- person-researcher agent for sections A-B (writes directly to document)
🎯 COMPLETED: [SKILL:person-osint] [profile of X completed]
🗣️ CUSTOM COMPLETED: [SKILL:person-osint] [Person profile ready]
