Manages learning extraction and sign application workflows. This skill should be used when extracting learnings from session transcripts, reviewing proposed signs, or applying approved learnings to CLAUDE.md files. Triggers on "extract learnings", "review signs", "apply signs", "manage signs", "learning backlog".
From m42-signsnpx claudepluginhub mission42-ai/m42-claude-plugins --plugin m42-signsThis skill uses the workspace's default tool permissions.
assets/backlog-template.yamlassets/learning-extraction-workflow.yamlassets/sprint-with-learning.yamlreferences/backlog-schema.mdreferences/claude-md-format.mdreferences/transcript-format.mdGuides Next.js Cache Components and Partial Prerendering (PPR) with cacheComponents enabled. Implements 'use cache', cacheLife(), cacheTag(), revalidateTag(), static/dynamic optimization, and cache debugging.
Migrates code, prompts, and API calls from Claude Sonnet 4.0/4.5 or Opus 4.1 to Opus 4.5, updating model strings on Anthropic, AWS, GCP, Azure platforms.
Details PluginEval's skill quality evaluation: 3 layers (static, LLM judge), 10 dimensions, rubrics, formulas, anti-patterns, badges. Use to interpret scores, improve triggering, calibrate thresholds.
This skill provides workflows for the learning loop - extracting wisdom from session failures and applying them as permanent signs in CLAUDE.md files.
See the plugin commands for available operations:
/m42-signs:add - Manually add a learning to the backlog/m42-signs:list - List learnings in the backlog/m42-signs:extract - Extract learnings from a session transcript/m42-signs:review - Review and approve/reject learnings/m42-signs:apply - Apply approved learnings to CLAUDE.md filesIntegrate learning extraction into your sprint workflows for automated capture of development insights.
During sprint execution, session transcripts are stored at:
.claude/sprints/<sprint-id>/transcripts/<phase>-<step>-<subphase>.jsonl
| Variable | Available In | Description |
|---|---|---|
{{sprint.id}} | All phases | Sprint identifier (e.g., 2026-01-18_feature-name) |
{{sprint.name}} | All phases | Human-readable sprint name |
{{step.id}} | For-each phases | Step identifier (e.g., step-1) |
{{step.prompt}} | For-each phases | Step description from SPRINT.yaml |
{{phase.id}} | All phases | Current phase identifier |
Add as a final phase in any sprint workflow:
phases:
# ... your existing phases ...
- id: learning-extraction
prompt: |
Extract learnings from this sprint's development sessions.
Use /m42-signs:extract to process transcripts at:
.claude/sprints/{{sprint.id}}/transcripts/
Use --confidence-min medium to filter low-confidence patterns.
Two templates are provided in assets/:
learning-extraction-workflow.yaml - Standalone learning extraction workflow
# Reference as a workflow
- id: learning
workflow: learning-extraction-workflow
sprint-with-learning.yaml - Complete sprint workflow with learning extraction
# Use as your sprint workflow
workflow: sprint-with-learning
Full sprint configuration with learning extraction:
# SPRINT.yaml
name: Feature Implementation Sprint
workflow: sprint-with-learning
steps:
- id: step-1
prompt: Implement user authentication
- id: step-2
prompt: Add authorization middleware
# Learning extraction runs automatically after all steps
To manually extract learnings from a specific transcript:
/m42-signs:extract .claude/sprints/<sprint-id>/transcripts/<file>.jsonl --confidence-min medium
After extraction completes:
/m42-signs:status to see backlog summary/m42-signs:review to approve/reject learnings/m42-signs:apply to apply approved signs to CLAUDE.md files