don't have the plugin yet? install it then click "run inline in claude" again.
+trigger phrases are reasonably specific and varied across report types (weekly, monthly, project, meeting notes)
+output format example shows concrete structure with sections for completed work, metrics, next steps, risks
weaknesses
~procedure section describes 'what' (ai understanding, markdown formatting) not 'how' - no step-by-step workflow or input/output contracts per step
~no failure modes documented: what happens if lark api fails, if conversation history is incomplete, if template is malformed, or if user edits conflict with concurrent access
~missing decision points - no mention of how the skill validates inputs, selects templates, or determines report scope when context is ambiguous