DESIGN.md@skill-evolutionA
DESIGN.md@skill-evolution from labs21-dev/agents-stack is a DESIGN.md file: the design rules coding agents follow in that repo.
mdr add labs21-dev/agents-stack/skill-evolution/DESIGN.md@git:20260912.9d47cf0Pins git:20260912.9d47cf0 in mdr.lock. Pin the fingerprint instead.
- Audit grade
- A, 14 of 14 checks passed
- Kind
- DESIGN.md file
- Repository
- labs21-dev/agents-stack
- Path
skill-evolution/DESIGN.md- Versions recorded
- 1
- Latest version
- git:20260912.9d47cf0, committed Sep 12, 2026
- Size
- 9,994 bytes
- Repository stars
- 27
- License
- none declared
Collected from public GitHub. Every version is saved exactly as it was and never changes. Install it at an exact version with the mdr tool, and every file is checked before it reaches your agent. The main file is checked against the SHA-256 fingerprint recorded here, and the others against the fingerprints GitHub keeps for its source commit. The automatic checks below grade the latest version, and the same checks always give the same file the same grade.
What the file says
# Design — skill-evolution（meta-skill 進化機制） > 狀態：設計定稿待批准（2026-07-22）。批准後才實作 scaffold 檔。 > 母分析：本對話以 meta-thinking-framework v2 跑的「meta-skill 系統元素與設計」分析。 > 起點：方案 B（扁平 skill-creator + 可選演化套件）→ 目標演化向 A（完整四層 stack）。 ## 0. 目標與邊界 - **目標**：給任何 skill 一個可選附掛的進化能力——收集案例 → 累積證據 → 達門檻 → AI 提議升格 patch → 人類批准 → 升格（對稱下架）。作為 `skill-evolution` skill + 可複製模板雙軌。 - **邊界**： - 不修改官方 `skill-creator` 本體（外部依賴，plugin marketplace）。 - 不碰 skill-creator 的 eval loop——進化層附加其上，eval 仍人類驅動。 - 對齊 ARC `mechanism not policy`：提供機制，skill 自決是否掛載。 - **可逆性**：中。單 skill 可回退；但 meta-criteria 若錯，汙染面廣 → 偏好可選附掛、保守升格、before-snapshot 可回退。 ## 1. 四層模型（B 起點做到哪、A 目標補哪） | 層 | 職責 | B 階段 | A 目標 | |---|---|---|---| | **L0 生成** | intent → skill draft | 借官方 skill-creator，不重造 | 同 | | **L1 評估** | 量測 skill 品質 | 借官方 skill-creator eval；**加指標↔真實案例雙向校準** | 完整雙向校準機制 | | **L2 演化** | 案例 → 升格/下架 | **本 skill 主體** | 同 + 更強 patch 工具 | | **L3 元標準** | 定義「什麼是好 skill」 | **輕量 accumulate-only**（compound/ 收案例，人類獨佔立法權，只 accumulate 不 legislate） | 跨 skill 收斂 N≥3 → 人類立法折進 constitution | > B 階段不強行做完整 L3——反身性鏡頭指出 L3 必須有外部 anchor，而外部 anchor 要靠跨 skill 案例累積才能形成。現在硬建 L3 = 自生標準 = 反身性失敗。故 B 階段只 accumulate，案例夠了再正式化立法。 ## 2. 反身性防禦（結構必要，非偏好） …
Read the whole file at its exact version.
How to install
mdr add labs21-dev/agents-stack/skill-evolution/DESIGN.md@sha256:22344f737b012ae5The line at the top of this page pins the label git:20260912.9d47cf0, which is what most people want. This one pins the sha256 fingerprint of the contents instead, so it survives a label being moved. Either way the exact version is written to mdr.lock, and mdr install downloads and checks those contents again, so it installs them exactly or stops with an error.
[](https://markdownregistry.com/a/art_6au3w37nai7ndf77)
5 badge image fetches in 30 days. What the badge shows.
Versions
The registry has recorded 1 version of this file, committed on Sep 12, 2026.
Audit of the latest version
- pass: Size between 200 bytes and 200 KB (9994 bytes)
- pass: No zero-width or bidi control characters
- pass: No instruction hidden inside an HTML comment
- pass: No link to an exfiltration or paste host
- pass: No credential-shaped string
- pass: No instruction to send local credentials anywhere
- pass: No text hidden with inline styles
- pass: No prompt-injection phrasing
- pass: No curl or wget piped into a shell
- pass: No recursive delete of root, home or parent
- pass: No instruction to read or print local credentials
- pass: No base64 blob over 200 characters
- pass: No link to a raw IP address
- pass: No script tag
Source
labs21-dev/agents-stack · 27 stars · license none declared · pushed 2026-09-19 · branch main
API
GET https://markdownregistry.com/api/v1/artifacts/art_6au3w37nai7ndf77 GET https://markdownregistry.com/api/v1/resolve?ref=labs21-dev/agents-stack/skill-evolution/DESIGN.md GET https://markdownregistry.com/api/v1/blob/22344f737b012ae5b7425a48fe12bec80becc84dd6e3690aa3c6fe71f28b38fe
Your AI assistant does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.
Guides: what DESIGN.md is and how agents use it · which agent file does what.