Lead free-form implementation work through bounded delegation: define the exact outcome, launch the smallest useful executor team, integrate the result, and prove completion. Use when the user asks Work to build, fix, implement, refactor, or change something without selecting a focused Work skill.
Evidence-first, read-only audits that always use one gpt-5.6-sol xhigh analyst and return a self-contained fix plan suitable for a lower-capability execution model. Use when the user asks to audit code, a diff, a branch, an artifact, or a runtime flow.
Turn the current conversation into a concise, source-traceable checklist of deliverables, decisions, experiments, guardrails, or concerns, then optionally register one goal and execute every item. Use when the user wants to capture an ideated direction, plan, requirements, discussion, or implementation concerns as an…
Read-only code-simplification analysis that always uses one gpt-5.6-sol xhigh analyst and returns a self-contained cleanup plan suitable for a lower-capability execution model. Use when the user asks to clean code, remove smells, reduce overengineering, refactor helpers, or assess package choices.
Inspect the current implementation and surface a short list of concrete, evidence-backed concerns without editing or producing a full audit or fix plan. Use when the user asks what gives you pause, what might be wrong, or wants a lightweight implementation sanity check.
Show the separate Work skills and their purpose without running a task. Use when the user wants to remember how Work is organized or asks what Work skills are available.
Run a rigorous, one-question-at-a-time ideation session that turns a vague desire into a sharp problem, genuinely distinct options, and a testable next move. Use when the user wants to brainstorm, refine, compare, or pressure-test a product, feature, business, creative, architecture, or project idea with a candid…
Guide and verify the smallest useful new-project foundation by resolving essential product, platform, toolchain, dependency, and validation choices. Use when the user wants to initialize, scaffold, or start a project.
Turn a task into a detailed, read-only execution plan by always launching one gpt-5.6-sol xhigh analyst and producing a self-contained handoff suitable for a lower-capability execution model. Use when the user wants a plan before implementation.
Pass the Work golden rules directly into the current conversation without spawning agents, editing files, or changing project state. Use when the user asks for the quality bar, taste principles, or a ready-to-paste software standard.
Show a read-only snapshot of the current Work goal, checklist, agents, evidence, blockers, current step, and next parent action. Use when the user asks for Work status.
Inspect a codebase and write a small number of high-value tests for real, observable behavior using the repository's existing framework and conventions. Use when the user asks to add, improve, or write meaningful tests without optimizing for coverage.