Show Model
Use when the human asks to see the accepted model, during onboarding wrap-up when a viewer exists, or after an accepted package changes what should be shown.
Interaction Contract
Use for onboarding Block C or any owner-requested rhythm change. Records agent runtime interaction settings and schedules OpenClaw cron jobs; it does not write company-model cards.
Package Update
Use for weekly package release checks or an owner-approved package update request. Updates package code only; never migrates the accepted model automatically.
Drift Sweep
Use when cards reach next-audit. Checks accepted cards against read-only sources, routes mismatches to drift-flag, and reports what needs review.
Build Brain
Use after accepted ontology cards change. Compiles cards into the registry graph, checks links, and prepares the model for dashboard, MCP, or agent consumers.
Connect Source
Use before mining a new source. Registers a chat export, spreadsheet, PDF, repo, CRM, dashboard, or feed in 02-source-map.md with trust and read-only access policy.
Interpret
Use when answering questions about a module, term, owner, source, or metric from the accepted ontology. Read-only; flags low confidence and drift.
Promote Digest
Use when staged proposals need human review. Builds an accept/edit/reject digest with diff, source, impact, and validator status. The agent never promotes.
Synthesize Digest
Use for a daily or weekly ontology digest. Summarizes accepted changes, pending proposals, drift, gaps, metric movement, and open human requests.
Business Ontology
Agent skill and reference runtime for maintaining a git-backed business ontology: source intake, review gates, evals, and MCP/GBrain boundaries.
Drift Flag
Use when an accepted card no longer matches reality. Records the mismatch in 08-drift-and-open-questions.md and routes fixes through propose-change.
Meeting Transcript Ingest
Use when a meeting transcript packet is ready and should become source events, model-change packages, human requests, and a digest without accepted model writes.
Mine Materials
Use when raw materials should populate the ontology. Extracts candidate concepts, modules, systems, interfaces, states, and decisions for human approval.
Grill Gaps
Use when open ontology gaps block a decision. Asks one focused question at a time, with a recommended answer and a fixed question budget.
Decide Like Module
Use when asked how a module should handle a borderline case. Reads accepted decisions and rules, answers with citations, and escalates when the model is silent or conflicting.
System Analysis
Use when a systems-thinking skill needs a bounded, source-backed projection from the accepted business ontology: workflows, states, metrics, decisions, constraints, and drift.
Daily Ingest
Use when a daily source packet from chat exports needs agent interpretation into source events, model-change packages, human requests, and a compact digest.
Propose Change
Use when a finding should change the ontology. Stages a proposal for human approval; never writes directly to accepted cards.
Extract From Input
Use when one untrusted input needs to become candidate ontology facts. Mines redacted facts with source locators and sends them to propose-change; never commits.
Onboard Contour
Use for the first-session Block A onboarding contour: a short ladder that frames the company, starting area, flow, source of truth, roles, and success metric before source setup.
Meeting Recorder
Use when a direct agent message or a group mention contains a Zoom, Google Meet, or Microsoft Teams link that should be recorded through the meeting recording runtime.
Business Ontology
Use when building, updating, auditing, or packaging a business ontology: definitions, states, decisions, sources, and drift for a module or company. Runs a live capture session, not general consulting.