Personal Assistant

Database: alh_personal · Repo: alhazen-skill-personal-assistant

The Personal Assistant application is an executive AI team — the digital staff a leader never had the headcount for. It is modelled on Nufar Gaspar’s “executive AI capability overhang”: a research analyst, a strategic thought partner, a communication expert, and an operational powerhouse, plus a career graph and a chief of staff who orchestrates them all. Every skill shares the alh_personal database, so the team works from one picture of your people, decisions, and work in flight.

The team

Skill	Role	What it does
Career	Career graph	People, collaborators, projects, and the opportunity pipeline — your career as a knowledge graph, not just a job search.
Research Analyst	Research	Frames decision-linked questions and commissions the Deep Research engines, then verifies and gates the findings before they touch a decision.
Board of Advisors	Strategic thought partner	A persona board debates a decision then converges — calibrated pushback, bias checks, scenario stress-tests, a decision journal.
Scribe	Communication	Writes in your voice for your audience — voice profile, reader personas, dimension-scored feedback.
Operations	Operational powerhouse	Recurring briefs, a stakeholder CRM, meeting prep, commitment tracking, and standing monitors.
Chief of Staff	Capstone	A read-only synoptic view — the daily agenda and weekly review across the whole team.
Health Coach	Health	Apple HealthKit ingestion, metric aggregates, goals, appointments, and regression alerts.

One database, shared entities

Because the team shares alh_personal, skills share data on purpose. People are the single alh-person entity, so a stakeholder tracked in Ops and a collaborator in Career are the same person. The Chief of Staff reads every teammate’s namespace (ops-, advsr-, scribe-, anlst-, career-) to build one agenda, and the Analyst feeds verified findings into decisions run by the Advisor. Select alh_personal in the dashboard switcher to see the whole team’s dashboards.

Operating principles

The team is built around five operating principles the skills enforce in common:

Speak, don’t type — every decision, piece, or opportunity opens with a messy voice-dictated primer brain dump; the agent structures it, never the operator.
Interview-first — the agent interviews you before it produces.
Plan / execute separation — framing is distinct from doing.
Deliberate intervention points — explicit ⏸ operator checkpoints; the agent works freely between them.
Wisdom of the crowd + a trust gate — consensus across advisors or research engines, then a sign-off gate before anything lands.

Reaches into Deep Research

The Personal Assistant commissions, it does not research. The Analyst dispatches questions to the Deep Research engines in alh_deep_research and reads their reports back, linking across the database boundary by reference. See the Multi-Database Model for how that works.