By name: learning-engineer description: Capture and structure lessons from discovery incidents, major insight reversals, or process failures. Use after meaningful mistakes, unexpected findings, or major workflow changes.
Learning Engineer
Capture reusable lessons to improve the discovery process over time.
Required Inputs
AGENTS.md- Description of the incident, reversal, or lesson
- Relevant artifacts (insight entries, story drafts, analyst notes)
Workflow
- Describe the event: what happened, when, in which discovery phase.
- Identify root cause: why it happened.
- State the impact: what had to be redone, what was delayed, what was incorrect.
- Write a prevention rule or process check.
- Classify: reusable across projects (promote candidate) or project-specific (keep local).
- Add entry to LEARNING_LOG.md.
Entry Format
## LEARN-[ID]: [Short Title]
- Date:
- Phase: [intake / clustering / journey / stories / PRD / roadmap / handoff]
- Event: [what happened]
- Cause: [why it happened]
- Impact: [what had to change]
- Prevention: [rule or checklist item to prevent recurrence]
- Promote candidate: [yes / no]
Required Output
- New entry in
LEARNING_LOG.md. - If promote candidate: draft reusable rule for canonical skill file.
Constraints
- Do not skip learning entries for major reversals.
- Do not include project-identifying details in promote candidates.