skill-agent-planning

Planning new skills or agents requires deciding whether to deliver as a skill, an agent, or both, then enumerating integration tasks, stakeholders, and artifacts required to ensure end-to-end readiness.

You should deliver as a skill, an agent, or both based on the feature's integration requirements and governance considerations. Use delivery vehicle decision guidance to evaluate reference patterns and choose the appropriate architecture for your workflow.

Integration tasks for a new skill include aligning planning outputs with artifact contracts, configuring validation hooks, updating rosters, and completing integration checklists to coordinate with existing factory configuration and avoid poorly integrated changes.

To validate new agents, align planning outputs with validation hooks and artifact contracts during the decomposition phase. This strategy ensures end-to-end readiness and coordinates with existing factory configuration to maintain governance standards.

Yes, you can use reference patterns and integration checklists to coordinate skill updates with existing factory configuration. These governance considerations prevent poorly integrated changes by aligning validation hooks and roster updates with artifact contracts.

When decomposing a feature request for a new agent, you need to enumerate the specific artifacts required for delivery, align them with artifact contracts, and ensure they support validation hooks and roster updates for complete integration readiness.