Skip to content
Back to skills

Cocowisdom

ASecurity

Governed CocoWisdom reflection lifecycle with read-only proposals, authorship checks, evidence ledgers, probation, and adherence-based archival.

  • 724 stars
  • 0 votes
  • 0 copies
  • 1 view
  • Added October 1, 2026
ai-agentsgo

Security analysis

A100/100

Pro scans all 20 files and shows the line behind each finding

Scanned October 1, 2026

npx -y skills add Snowflake-Labs/cocoplus --skill cocowisdom --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Cocowisdom?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Cocowisdom
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/snowflake-labs-cocowisdom-bedc27b9/badge)](https://www.skillsdirectory.com/skills/snowflake-labs-cocowisdom-bedc27b9)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: "wisdom-reflect"
description: "Governed CocoWisdom reflection lifecycle with read-only proposals, authorship checks, evidence ledgers, probation, and adherence-based archival."
version: "2.0.2"
author: "CocoPlus"
tags:
  - cocoplus
  - cocowisdom
  - reflection
user-invocable: false
blocking: false
---

## Objective

At each configured reflection boundary, route the queued context to the read-only `coco-wisdom-reflector` agent. Its output is a proposal only. Before any write, validate structural completeness, safety, and autonomous authorship; then route accepted intents through the deterministic promoter.

## Required Contract

- Accept only `add`, `merge`, `patch`, `drop-support-file`, or `delete` intents with reason and evidence.
- Reject targets authored by a human or installed from a marketplace.
- Append every accepted write to the pattern's `.ledger.jsonl` with a content-addressed evidence reference.
- Recall probationary patterns normally, but exclude them from capacity and archival until the tier maturity gate is met.
- Measure survival by calls over relevant requests. Never use wall-clock age.
- Archive the lowest-adherence mature patterns when the project or global cap is exceeded. Never delete pattern history.

## Exit Criteria

- [ ] The reflector has no write tools.
- [ ] Every mutation passes the promoter gate and creates a ledger entry.
- [ ] Project and global tiers use separate maturity gates and capacity caps.

## Anti-Rationalization

| Shortcut / Temptation | Why It Fails |
|---|---|
| Let the reflector apply its own proposal | Proposer/writer separation is the core safety boundary. |
| Archive based on age | Seasonal patterns must be judged only when opportunities arise. |

Files in this skill

  • wisdom-consolidation-log.skill.md1.3 KB
  • wisdom-distill.skill.md3.4 KB
  • wisdom-forget.skill.md1.3 KB
  • wisdom-get.skill.md2.1 KB
  • wisdom-index.skill.md1.9 KB
  • wisdom-insights.skill.md4.7 KB
  • wisdom-keep.skill.md1.3 KB
  • wisdom-learn.skill.md1.3 KB
  • wisdom-learnings.skill.md1.2 KB
  • wisdom-list.skill.md2.5 KB
  • wisdom-recall.skill.md1.8 KB
  • wisdom-reflect.skill.md1.7 KB
  • wisdom-reject.skill.md1.8 KB
  • wisdom-review.skill.md1.7 KB
  • wisdom-route.skill.md1.5 KB
  • wisdom-search.skill.md2 KB
  • wisdom-status.skill.md1.3 KB
  • wisdom-writer.skill.md1.5 KB
  • wisdom.skill.md6.8 KB

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…