What problem does it solve?
AI agents often ask users "who is X?" when the answer already exists in the knowledge base, wasting user attention on questions the system can answer itself. This Skill gates every identity, role, or relationship question behind a mandatory lookup chain so the user is only bothered as a last resort.
Core Features & Use Cases
- Ordered Lookup Chain: Runs a six-step escalation — think/synthesize, search plus full page read, mounted source queries, timeline and graph traversal, web search, and only then user escalation — stopping at the first clear answer.
- Hypothesis-Led Escalation: When asking is justified, the question leads with a best-guess hypothesis, states what was searched and found, and uses a 2-4 option choice gate instead of a bare "who is X?".
- No-Placeholders-at-Ingest Rule: Post-ingestion passes scan newly created person and company pages for placeholder text like "[To be filled by content analysis]" and resolve relationship and role fields immediately via put_page.
- Use Case: During a bulk email import that creates 100+ contact pages, run this Skill to fill each page's relationship field from email domains, meeting history, and timeline signals, so no briefing ever shows an "unknown relationship" entry.
Quick Start
Before asking me who a contact is, run the resolve-before-asking chain to check the brain's pages, timeline, and sources, and only escalate with a hypothesis if nothing resolves.
