What problem does it solve?
This skill closes the feedback loop in PM OS by enabling structured logging of decisions, post-launch reviews, and learning patterns from history to continuously improve planning and execution.
Core Features & Use Cases
- Unified learning modes: --decision to log decision context with rationale, --launch to run post-launch retros, and --patterns to extract recurring rules from history.
- Output-driven memory: writes decision logs to outputs/decisions, launch reviews to outputs/reviews, and learned rules to .claude/rules/learned for future automation.
- Use Case: After a major initiative, run --decision to capture context, later run --launch for a post-mortem, and periodically run --patterns to refine playbooks.
Quick Start
Run the learning commands to begin capturing decisions, launches, and patterns:
- pm-os learn --decision
- pm-os learn --launch
- pm-os learn --patterns