geas

Turns a Geas User goal into reviewable Mission Spec, Mission Design, Pre-build Design Surface decisions, and initial Task Contract Set baselines. Use during Mission intake, baseline revision, scope clarification, acceptance criteria definition, implementation-before-build decision surfacing, task slicing, or building-entry readiness. Do not use for Task implementation, role Evidence, verification, review, challenge Evidence, or Mission closure.

Coordinates Geas Mission work as the user entrypoint start a Mission, recover current state, choose specifying/building/consolidating stage handoffs, and prepare User briefings. Use when the user asks to run, continue, inspect, close, or route a Geas Mission. Do not use for writing role Evidence or performing Task implementation, verification, review, or challenge work.

Coordinates Geas Task execution loops from accepted Task Contract through role handoffs, Task result User Judgment input, Task Evidence, and next Task or consolidation routing. Use when a Mission is in building stage, Task State needs phase dispatch or recovery, Task result briefing is needed, or accepted Task results must be recorded. Do not use for writing implementation, verification, review, challenger Evidence, Mission baseline drafts, Mission closure, or User Judgment decisions.

Pressure-tests Mission baseline candidates or Task results for hidden assumptions, scope boundary problems, verification gaps, operational risks, tradeoffs, and repeat-risk patterns. Use when a coordinator or stage handoff asks for challenge before baseline acceptance or Task result judgment, especially after Implementation, Verification, and Review Evidence exist. Do not use for implementation fixes, verification checks, review findings, Task Evidence, User Judgment, Mission Evidence, or direct artifact revision.

Synthesizes accepted Task Evidence against the Mission baseline, prepares Mission result User Judgment input, and records Mission closure artifacts after User Judgment. Use when a Mission is ready for consolidation, accepted Task Evidence must be compared to Mission Spec and Mission Design, or Debt, Memory, and Mission Evidence closure records need guarded coordination. Do not use for Task implementation, role Evidence, Task Evidence, baseline drafting, deciding User Judgment, or closing a Mission without User acceptance.

Provides the Geas runtime write adapter and status interface through the bundled `scripts/geas` CLI. Use when a Skill needs to initialize `.geas/`, read status, record Mission or Task artifacts, record User Judgment, transition state, or preserve payloads after CLI write failure. Do not use to author artifact content or decide whether Evidence is sufficient.

Performs Task Contract-scoped implementation work and records Implementation Evidence. Use when a coordinator hands off implementation phase work, code or document outputs must be changed inside an accepted Task Contract, and changed outputs, self checks, limits, or contract deltas must be returned. Do not use for Verification Evidence, Review Evidence, Challenger Evidence, Task Evidence, User Judgment, Mission baseline drafting, or Mission closure.

Performs Task Contract-scoped review of changed outputs and role Evidence, then records Review Evidence. Use when a coordinator hands off review phase work with a current Task Contract, Implementation Evidence, Verification Evidence, target outputs, and review focus; inspect quality, fit, boundaries, risks, Evidence sufficiency, findings, verdict, and recommendation. Do not use for implementation fixes, Verification Evidence, Challenger Evidence, Task Evidence, User Judgment, Mission baseline drafting, or Mission closure.

Performs Task Contract-scoped verification of changed outputs and records Verification Evidence. Use when a coordinator hands off verification phase work with a current Task Contract, Implementation Evidence, target outputs, and verification environment; run available checks, map results to acceptance criteria, expose unverified scope, and return a verdict. Do not use for implementation changes, Review Evidence, Challenger Evidence, Task Evidence, User Judgment, Mission baseline drafting, or Mission closure.