continuous-learning-v2

基于本能的学习系统，通过钩子观察会话，创建带置信度评分的原子本能，并将其进化为技能/命令/代理。v2.1版本增加了项目范围的本能，以防止跨项目污染。

Quality

61%

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 ./docs/zh-CN/skills/continuous-learning-v2/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 highly actionable with concrete commands, configs, and a clear workflow diagram, but it is verbose for a SKILL.md — duplicating the command list and carrying changelog/marketing/link padding — and keeps all reference material inlined rather than splitting it into separate files.

Suggestions

Remove the duplicate command listing: keep only the 命令 table and drop the redundant bash block in quick-start step 3 (or vice versa).

Move the two version-changelog tables, the file-structure tree, the config-key reference, and the scope-decision guide into a separate references/ file (e.g. REFERENCE.md) linked from a short overview, cutting the inlined bulk.

Trim the marketing intro, the "相关链接" external-link section, and the "为什么用钩子而非技能" rationale to the essential operational points.

Dimension	Reasoning	Score
Conciseness	Most content is operationally useful, but the 355-line body includes clear padding: the command set is listed twice (quick-start bash block and the 命令 table), two version-changelog tables (v1→v2, v2.0→v2.1), a marketing-style intro, and an external-links section that could be trimmed.	3 / 5
Actionability	It provides concrete, copy-pasteable guidance — real CLI calls ("python3 instinct-cli.py promote prefer-explicit-errors", "--dry-run"), slash commands, JSON hook/config blocks, a YAML instinct example, and an explicit directory tree — with only minor gaps in specifying command output/behavior.	4 / 5
Workflow Clarity	The pipeline is laid out clearly via the ASCII workflow diagram (session → hook capture → observer → pattern detection → create/update → /evolve + /promote) and a numbered quick-start, and promotion offers a --dry-run checkpoint; minor validation gaps remain (e.g. no explicit feedback loop for /evolve).	4 / 5
Progressive Disclosure	Section headers are well-organized, but at 355 lines everything is inlined into SKILL.md with no one-level-deep references to separate files; material like the file-structure tree, config reference, and scope-decision guide are candidates for split reference docs, and no bundle reference files are provided.	3 / 5
	Total	14 / 20 Passed

Description

58%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 gives a clear, specific account of what the system does but omits any explicit "when to use" trigger guidance and relies on technical jargon rather than natural user keywords. It is distinct as a niche but would benefit from a trigger clause.

Suggestions

Append an explicit trigger clause, e.g. "Use when setting up automatic learning from Claude Code sessions, configuring hook-based behavior extraction, or managing/promoting learned instincts."

Add natural-language keywords and synonyms users would actually say (e.g. "learn from sessions", "remember my patterns", "export/import learned behaviors") alongside the technical terms.

Mention the supporting commands (export/import/promote/projects) briefly so the capability set reads as comprehensive.

Dimension	Reasoning	Score
Specificity	The description names the domain and lists several concrete actions — "通过钩子观察会话" (observe sessions via hooks), "创建带置信度评分的原子本能" (create instincts with confidence scores), "进化为技能/命令/代理" (evolve into skills/commands/agents) — with only minor coverage gaps (e.g. export/import and promote are absent).	4 / 5
Completeness	It clearly answers "what" (observe → create → evolve instincts) but provides no "Use when..." trigger clause, so per the guideline an absent explicit trigger caps completeness at 3.	3 / 5
Trigger Term Quality	It contains relevant keywords (学习系统, 钩子/hooks, 本能/instincts, 置信度/confidence, 技能/命令/代理) but the terms lean technical/jargon and miss natural user phrasings, synonyms, and file extensions that a user would actually say.	3 / 5
Distinctiveness Conflict Risk	The instinct-based continuous-learning niche is fairly distinct (observe → score → evolve pipeline, project vs global scoping), with only minor overlap risk against general learning/memory skills.	4 / 5
	Total	14 / 20 Passed

Validation

93%

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 — 15 / 16 Passed

Validation for skill structure

Criteria	Description	Result
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	15 / 16 Passed

Repository: affaan-m/ECC
Path: docs/zh-CN/skills/continuous-learning-v2/SKILL.md
Commit: e04ea0b

Reviewed: 2 days 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.