reflect-on-work

Post-task reflections systematically document what worked, what failed, and patterns noticed after task execution. This Skill generates structured reflections in compact JSON format to ensure consistency across all agent outputs and create a foundation for knowledge updates that inform system evolution.

Effective reflections capture three core elements: what worked (successful approaches and decisions), what failed (obstacles and ineffective strategies), and patterns noticed (recurring themes across executions). This structure ensures reflections are honest, actionable, and feed directly into knowledge base updates.

Yes. This Skill applies mandatory post-task reflections across all domains and task types, generating structured outputs automatically after each execution. The consistent JSON schema ensures reflections scale across diverse workflows without manual formatting.

Reflections feed knowledge_updates fields that capture learnings for your knowledge base. By tracking what worked and what failed systematically, reflections create a compounding record that informs agent behavior refinement and prevents repeated mistakes across future executions.

Reflections conform to a defined JSON schema with dedicated fields for what_worked, what_failed, and patterns_noticed, plus knowledge_updates sections. This compact structure keeps reflections machine-readable and easily parsed downstream without sacrificing detail.