cm - Manage procedural memory for coding agents
Stores, retrieves, validates, and curates cross-agent coding knowledge from session history using playbooks, feedback, and confidence decay.
Tags
Updated: 2026-10-06Capabilities
Typical Inputs
Typical Outputs
What this skill does
- Retrieve task-relevant rules
- Search related playbook rules
- Record rule feedback
- Validate proposed rules
- Extract rules from sessions
- Audit sessions against rules
- Track rule confidence decay
- Deprecate obsolete rules
- Import and export playbooks
- Run system diagnostics
- Analyze memory gaps
- Show usage statistics
Inputs
- Task description
- Session transcripts
- Rule identifiers
- Proposed rule text
- Rule feedback
- Feedback reasons
- Time ranges
- Playbook files
- Starter name
Outputs
- Relevant rules
- Related diary entries
- Gap analysis
- Rule evidence chains
- Validation recommendations
- Updated playbook rules
- Session audit results
- Confidence scores
- Exported playbook files
- System health reports
- Usage statistics
Requirements
- Accessible cass installation
- Access to session history
- Playbook storage access
- LLM support for extraction stages