Codify Pulse work learnings into memory artifacts for future planning.
npx skills add https://github.com/quanpersie2001/pulse --skill compounding
Please help me install this Agent Skill. Skill: compounding Source: https://github.com/quanpersie2001/pulse/tree/main/plugins/pulse/skills/compounding Command: npx skills add https://github.com/quanpersie2001/pulse --skill compounding
Capture durable learnings from completed Pulse work so future planning gets smarter and future beads carry the right context. Invoke after reviewing completes and the feature is merged, or after a debugging session uncovers a non-obvious root cause. Runs three parallel analysis subagents (patterns/decisions/failures), synthesizes into .pulse/memory/learnings/YYYYMMDD-<slug>.md, promotes critical items to .pulse/memory/critical-patterns.md. Classifies the rest for planner-only or bead-local propagation. Key output: critical-patterns.md is read by every planning and exploring Phase 0 — this is the flywheel that makes the ecosystem smarter over time.
Invoke compounding after review and merge to synthesize learnings into memory files and update the critical-patterns list.
High-intent search queries and answers about installing and using this skill.
You extract reusable learnings by running parallel analysis subagents for patterns, decisions, and failures, synthesizing the findings into durable memory artifacts to improve future planning cycles.
The best way to capture debugging root causes is to invoke a compounding process after the session, which codifies the non-obvious root cause into a memory artifact under the learnings directory.
Promoting critical patterns to a shared file makes the ecosystem smarter because every planning and exploring Phase 0 reads this list, creating a flywheel that continuously feeds historical context into new workflows.
You should run memory compounding immediately after reviewing completes and the feature is merged, ensuring that bead-local or planner-wide memory is updated before the next planning phase begins.
Yes, the workflow requires the beads-cli dependency to function, as it relies on the underlying bead architecture to propagate classified insights into either bead-local or planner-wide memory.
Limitations depend on propagation rules: only critical items are promoted to planner-wide memory, while the rest are classified as bead-local context, restricting broader ecosystem visibility for non-critical findings.
quanpersie2001Triage and repair blocked builds, failing tests, and runtime crashes in Pulse projects.
code-dot-orgGenerate implementation plan design artifacts from a plan template.
bacchus-labsCoordinate parallel subagents to refresh governance documents and reconcile issues.
5kahoisaacStore and recall persistent AI agent memories using markdown files with YAML frontmatter.
