self-evolution

Scope Guard + Process Evolution + Knowledge Evolution — 主动护栏与自我进化。 Use when: operator scope 发散偏离愿景、同类错误反复出现、SOP 流程缺口、有价值的知识/方法论值得沉淀。 Not for: 日常 SOP 推进（正常执行）、一次性个案 bug fix。 Output: Scope Guard Log 记录 / Evolution Proposal 提案 / Episode Card → Method/Skill 蒸馏 → Eval 验证。

Quality

70%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./cat-cafe-skills/self-evolution/SKILL.md

Quality

Content

63%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

The body is well-organized with clear per-mode workflows, guardrails, and tabular reference material, yielding good conciseness and workflow clarity. Its weaknesses are actionability (process prose over executable examples) and progressive disclosure (many cited external files are not present in the bundle, leaving a monolithic single-file skill).

Suggestions

Add at least one fully-worked inline example per mode (e.g. a sample filled Scope Guard Log row, a sample EP-XXX proposal, a sample Episode Card) so the guidance is copy-paste actionable rather than template-naming only.

Either ship the referenced bundle files (TEMPLATE.md for episodes/methods/evolution-proposals, evals/mode-c/) or remove the dangling path references so progressive disclosure is real, not promised.

Trim the revision-history asides (e.g. the '2026-07-15 修订' parenthetical and F086/F167 lesson cross-references) into a short changelog or deprecation note to recover tokens.

Dimension	Reasoning	Score
Conciseness	The body is dense and table-driven with little padding or explanation of concepts Claude already knows; a few sections (e.g. the 理解偏差自记录 backstory and F167 revision note) could be trimmed slightly.	4 / 5
Actionability	Provides concrete templates, log formats, and 5-slot structures, but much guidance is process description rather than executable commands/code; the referenced templates (docs/episodes/TEMPLATE.md, evals/mode-c/TEMPLATE/) are named but not inlined or exemplified.	3 / 5
Workflow Clarity	Each mode has a clear numbered sequence with explicit checkpoints ('先闭环当前任务', '30 天验证', escalation/approval gates, hard guardrails), satisfying most of the workflow anchor with only minor validation gaps.	4 / 5
Progressive Disclosure	References several external paths (docs/scope-guard-log.md, docs/evolution-proposals/TEMPLATE.md, docs/methods/TEMPLATE.md, evals/mode-c/) but no bundle files actually exist, so the body is effectively a single monolithic file with inline references that cannot be navigated.	3 / 5
	Total	14 / 20 Passed

Description

76%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

The description is well-structured with explicit what/when/not-for/output clauses and concrete output artifacts, giving it strong completeness and specificity. Its main weakness is trigger-term quality: the phrasing relies on internal project vocabulary ('operator scope 发散') rather than language a user would naturally utter.

Suggestions

Rewrite the 'Use when' trigger phrases in plain user-facing terms (e.g. 'when the conversation drifts away from the current task's goal') rather than 'operator scope 发散偏离愿景'.

Add common synonyms / natural keywords to the triggers so the skill matches the way users actually phrase scope-creep or repeated-mistake situations.

Keep third-person voice consistent — the description is already third person, but ensure trigger phrasing reads as a user statement, not internal team shorthand.

Dimension	Reasoning	Score
Specificity	Names the three concrete modes (Scope Guard, Process Evolution, Knowledge Evolution) and lists several specific outputs (Scope Guard Log, Evolution Proposal, Episode Card, Method/Skill distillation, Eval), which is strong coverage with only minor abstraction around the mechanisms.	4 / 5
Completeness	Explicitly answers what (the three modes and their outputs) and when (a 'Use when:' clause, a 'Not for:' clause, and an 'Output:' line) with concrete trigger phrases, matching the top anchor.	5 / 5
Trigger Term Quality	Includes some natural trigger phrases ('同类错误反复出现', 'SOP 流程缺口', '一次性个案 bug fix') but leans heavily on project jargon ('operator scope 发散偏离愿景') that a user would not naturally say, missing common synonyms.	3 / 5
Distinctiveness Conflict Risk	Has a clear niche (meta-process/self-evolution) with explicit 'Not for' boundaries and a disambiguation section against sibling skills; minor overlap risk with collaborative-thinking/deep-research, which the body later addresses.	4 / 5
	Total	16 / 20 Passed

Validation

100%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: zts212653/clowder-ai
Path: cat-cafe-skills/self-evolution/SKILL.md
Commit: 090626a

Reviewed: ~23 hours ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.