name: persona-research
description: Public persona research authority — discover, rank, and curate public sources for a person across websites, YouTube, podcasts, newsletters, LinkedIn, and social profiles while avoiding name collisions, wrong-identity contamination, and low-signal pages
keywords: ["persona", "research", "public", "youtube", "linkedin", "authority", "discover", "rank", "curate", "sources", "person", "websites", "podcasts", "newsletters", "social"]
orb_class: trojan
persona-research
Use this when the task is to build a source base for a person-specific profile, partner agent, or high-fidelity voice model from public material.
Use It For
- Finding the right person across ambiguous search results
- Expanding from an official site into high-signal internal pages
- Ranking public sources by usefulness for persona modeling
- Separating self-authored material from reviews, testimonials, and commentary
Workflow
- Start with seeded facts:
- official website
- known YouTube/channel URLs
- LinkedIn/X/Instagram/newsletter/podcast URLs
- domain keywords and exclude keywords
- Discover candidate sources:
- search the person name alone
- search the person name with domain-specific terms
- search the official domain for bio, about, newsletter, podcast, articles, services
- search YouTube watch pages, not just channel pages
- Score and keep sources in this priority order:
- official bio/about pages
- self-authored newsletters and articles
- direct video/podcast episode pages with transcript potential
- self-authored landing pages describing methods, products, or worldview
- secondary profiles like LinkedIn
- De-prioritize or exclude:
- wrong-person matches
- generic channel pages
- widget-heavy social embeds
- testimonials and student/customer quotes
- pages dominated by site chrome rather than authored text
Source Quality Heuristics
- Prefer first-person language over third-party praise
- Prefer pages with concrete claims, methods, beliefs, and tradeoffs
- Prefer episode pages and transcripts over thumbnails, posts, or playlists
- Prefer stable identity pages for biography and claims
- Prefer public sources that can be cited later
Identity Collision Handling
When multiple people share the same name:
- add
focus_keywords tied to the target domain - add
exclude_keywords tied to the wrong identity - boost
preferred_domains - require at least one official domain or self-authored page before accepting a new cluster of sources
Output Shape
Produce a source set with:
- URL
- source type
- whether it appears self-authored
- whether it appears transcript-capable
- why it was kept
- why low-signal candidates were excluded
Rules
- Treat biography and identity claims as highest-risk; prefer official site or direct profile sources
- Treat testimonials as evidence of offer reception, not identity or worldview
- Treat social/channel pages as discovery aids unless they contain direct authored text or transcripts
_Last reviewed: 2026-05-14 — automated polish pass per issue #75._