Stats
Actions
Tags
From symbiopowers
Use after shipping a meaningful chunk of work, before merge, or when closing a major session to capture lessons in docs/solutions and propose manifest or rules updates as human-reviewed diffs only.
How this skill is triggered — by the user, by Claude, or both
Slash command
/symbiopowers:compound-retrospectiveThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
**Purpose:** Turn friction, surprises, and validator disagreements into **durable repo knowledge** — similar in spirit to compounding documentation in mature agent workflows, but with **strict human control** over policy files (`AGENTS.md`, `orchestrator-manifest.yaml`).
Purpose: Turn friction, surprises, and validator disagreements into durable repo knowledge — similar in spirit to compounding documentation in mature agent workflows, but with strict human control over policy files (AGENTS.md, orchestrator-manifest.yaml).
Core principle: Write solutions; propose policy diffs; never silently rewrite organizational rules.
Announce at start: "I'm using the compound-retrospective skill to record lessons learned."
docs/solutions/YYYY-MM-DD-<short-slug>.mdInclude:
| Section | Content |
|---|---|
| Context | Why this work happened |
| What we learned | Bullet list, concrete |
| Decisions | Tradeoffs, with rationale |
| Follow-ups | Optional tickets or future plan tasks |
Use relative links to specs/plans if relevant.
If rules should change:
AGENTS.md / orchestrator-manifest.yaml / checklistsDO
DON'T
AGENTS.md without reviewPairs with:
skills/finishing-a-development-branch — run this skill before or during branch completionskills/validation-aggregator — cite which gates failed and why useful rules emergednpx claudepluginhub vonschau/symbiopowers --plugin symbiopowersGuides creation, editing, and verification of skills for AI coding agents using test-driven development with subagent scenarios. Use when authoring or debugging skills.