How does the extraction process work?

The skill parses session transcripts stored in JSONL format from sprint directories, identifying recurring patterns or lessons learned during the development process.

Can I manually add learnings to the backlog?

Yes, you can use the /m42-signs:add command to manually insert specific insights into the learning backlog for later review and application.

What are 'signs' in the context of this Claude Code skill?

Signs are permanent learnings, patterns, or rules extracted from development sessions that are applied to CLAUDE.md files to guide Claude's future behavior and decision-making.

Does this skill automatically update my documentation?

The skill includes a controlled review process. You use /m42-signs:review to approve or reject proposed learnings before running /m42-signs:apply to update your CLAUDE.md files.

Sign Management Workflow

Name: Sign Management Workflow
Author: Mission42-ai

byMission42-ai

生产力与工作流

Automates the extraction of development insights from session transcripts and applies them as permanent guidance to CLAUDE.md files.

关于

This skill implements a continuous learning loop for developers by analyzing session transcripts to identify patterns, technical failures, and domain-specific wisdom. It allows users to manage a learning backlog, review proposed improvements, and automatically update CLAUDE.md files with refined technical signs or rules. By integrating directly into modular sprint workflows, it ensures that project-specific knowledge is captured and maintained, preventing the repetition of past mistakes and improving the overall efficiency of AI-assisted development.

主要功能

Seamless integration with sprint orchestration templates
0 GitHub stars
Learning backlog management with review and approval workflows
Automated application of approved learnings to CLAUDE.md
Automated learning extraction from session transcripts
Confidence-based filtering to ensure high-quality pattern capture

使用场景

Extracting technical debt lessons after a complex sprint implementation
Capturing institutional knowledge from session histories for better team onboarding
Maintaining project-specific coding standards and rules in CLAUDE.md

关于

主要功能

Seamless integration with sprint orchestration templates
0 GitHub stars
Learning backlog management with review and approval workflows
Automated application of approved learnings to CLAUDE.md
Automated learning extraction from session transcripts
Confidence-based filtering to ensure high-quality pattern capture

使用场景

Extracting technical debt lessons after a complex sprint implementation
Capturing institutional knowledge from session histories for better team onboarding
Maintaining project-specific coding standards and rules in CLAUDE.md