Distill terminal session outcomes into CocoWisdom using the project schema. Usage: $wisdom distill [session-id|run-id]
Scanned 9/5/2026
Install to Claude Code
npx -y skills add Snowflake-Labs/cocoplus --skill cocowisdom --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Cocowisdom?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/snowflake-labs-cocowisdom-9d58c402)More formats (shields.io, HTML) on the badges page.
---
name: wisdom-distill
description: "Distill terminal session outcomes into CocoWisdom using the project schema. Usage: $wisdom distill [session-id|run-id]"
version: "2.0.2"
author: "CocoPlus"
tags:
- cocoplus
---
# CocoWisdom Distill
Use this skill when `$wisdom distill` is invoked directly or when the Stop hook queues a background distillation request.
## Objective
Extract durable, reusable lessons from a completed, exited, or failed session and route them into CocoWisdom without blocking the developer's next turn.
## Inputs
- Optional session or run id from `$wisdom distill <id>`.
- `.cocoplus/wisdom/SCHEMA.md` when present.
- `.cocoplus/session/PROGRESS.md`, `.cocoplus/session/CONTEXT.md`, `.cocoplus/session/steps.jsonl`, and relevant lifecycle artifacts.
- `.cocoplus/wisdom/review-queue.md` for unresolved contradictions.
## Procedure
1. Check `.cocoplus/wisdom/SCHEMA.md`. If it does not exist, create a minimal schema with topic categories, source session id, outcome status, evidence pointer, confidence, contradiction handling fields, and an optional `## Stage Mappings` table for routing stage role patterns to positive wisdom topics.
2. Read only the terminal-session artifacts needed to identify lessons, rejections, corrections, governance decisions, and repeated failure modes.
3. Classify each candidate into a schema-declared category file. Do not invent standing-context text outside the schema.
4. If a candidate contradicts existing wisdom, follow `[wisdom].contradiction_action`:
- `queue`: add it to `.cocoplus/wisdom/review-queue.md`.
- `skip`: record the skipped contradiction in the report.
- `overwrite`: replace only when the evidence pointer and rationale are explicit.
5. Write accepted entries to the matching topic file under `.cocoplus/wisdom/`.
6. Write a distillation report to `.cocoplus/lifecycle/cocoflow/<run-id>/distillation-<timestamp>.json` when a run id exists, otherwise `.cocoplus/wisdom/distillation-<timestamp>.json`.
## Rules
- Distillation is asynchronous maintenance. It must never block Stop hook completion.
- Failed sessions are valid sources when `[wisdom].distill_on_failure = true`.
- Store evidence pointers, not transcript dumps.
- Preserve reviewability: every accepted or queued lesson needs source id, timestamp, category, and rationale.
- Prefer explicit curation for permanent memory. Automatic distillation may queue candidates, but durable project constraints, rejected approaches, and branch-scoped decisions need evidence pointers and reviewable rationale.
## Success Criteria
- [ ] Schema exists or is created.
- [ ] Accepted lessons are category-routed.
- [ ] Contradictions are queued or handled according to config.
- [ ] A distillation report records counts for accepted, queued, skipped, and errored candidates.
## Anti-Rationalization
| Shortcut / Temptation | Why It Fails |
|-----------------------|--------------|
| Treat the skill as complete because the file exists | Skill contracts must describe observable behavior and verification, not just command names. |
| Skip artifact and safety checks for a small command | Small commands still mutate state or guide execution; preserve the same gates. |
## Exit Criteria
- [ ] Command behavior matches the owning feature contract.
- [ ] Required reads, writes, and user-visible outputs are described.
- [ ] Safety, governance, and artifact constraints are preserved.
- [ ] Missing state produces a clear, non-destructive result.
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!