Mode Selection

Skill class selection

Classify the target skill before synthesis. This determines the coverage dimensions that must be represented in sources, references, and validation.

When the class is ambiguous, ask one direct clarification question before synthesis.

Execution shape selection

Choose the skill’s primary execution shape separately from the skill class. Class answers “what domain/problem is this skill for?” Shape answers “how should this skill run?”

Use references/execution-shapes.md for the full decision table and the next leaf reference to load.

Record:

1. primary execution shape
2. simpler-shape rejection when the chosen shape is advanced
3. exact leaf references opened because of that choice

Required outputs by path

• synthesis: source inventory, decisions, coverage matrix, gaps.
• synthesis: selected class, selected execution shape, and selected example profile path(s), including profile-requirement coverage.
• synthesis: simplicity rationale showing why the chosen shape is necessary and which simpler shapes were rejected.
• synthesis: portability note when provider-specific mechanics are used.
• synthesis: explicit retrieval stopping rationale showing why further collection is currently low-yield.
• authoring: updated SKILL.md and required supporting files.
• description optimization: should/should-not trigger sets and final description.
• iteration: example intake summary and behavior deltas.
• registration/validation: registration edits and validator results.

Completion Checks

Do not claim completion when any required path output is missing.

For authoring/generator skills, report missing transformed examples, selected-profile requirements, required class dimensions, or execution-shape choices as open gaps instead of hiding them. When advanced mechanics are used, include justification and portability notes or report the missing decision as an open gap.