continuous-learning

Continuous Learning

Purpose

Capture lessons learned from implementation, production, QA, and release so the project improves over time.

Use When

• a feature or milestone is complete
• the team repeatedly hits the same failure mode
• a postmortem should produce reusable knowledge

Inputs

• retrospective notes
• defect trends
• delivery outcomes
• tooling or workflow pain points

Process

1. collect what worked, what failed, and why
2. separate one-off events from repeatable patterns
3. extract improvements that should become rules, skills, or automation
4. assign owners for process changes
5. archive lessons in a searchable format

Outputs

• learning log
• candidate new skills or rules
• workflow changes
• retrospective summary

Quality Bar

• produces a current source of truth, not disconnected notes
• names owners, risks, and next actions explicitly
• separates decisions from assumptions and open questions

Common Failure Modes

• outdated docs that no longer match reality
• plans with no owner or no exit criteria
• hiding risks until they become schedule blockers

Related Agents

• producer
• doc-updater
• refactor-cleaner

Related Commands

• learn
• evolve
• update-docs

Notes

• Keep this skill aligned with the relevant rules layer and current project documentation.
• If engine-specific constraints materially change the workflow, hand off to the matching engine skill or engine-specific reviewer.