continuous-learning-v2

一种基于本能（Instinct）的训练系统，通过钩子（Hooks）观察会话，创建带有置信度评分的原子化本能，并将其进化为技能（Skills）、命令（Commands）或智能体（Agents）。

Quality

55%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./docs/ja-JP/skills/continuous-learning-v2/SKILL.md

Quality

Content

50%

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 and offers genuinely concrete commands and config, but it is verbose for a SKILL.md, keeps all reference material inline instead of splitting it out, and depends on an observe.sh script that is not shipped in the bundle.

Suggestions

Move stable reference material (file structure, scope-decision guide, version-comparison tables, confidence-score table) into separate reference files and link to them one level deep, keeping SKILL.md a lean overview.

Ship the referenced `observe.sh` (or correct the hook paths to a script that exists in the bundle) so the documented hook configs are actually executable.

Add explicit validation/checkpoint steps to the learning workflow — e.g., a verify step after observation analysis and before instinct creation, mirroring the existing --dry-run gate on promote.

Dimension	Reasoning	Score
Conciseness	The ~355-line body is mostly relevant operational guidance but includes lengthy conceptual material (version-comparison tables, an ASCII architecture diagram, the "why hooks vs skills" rationale, and a full scope-decision guide) that could be tightened.	2 / 3
Actionability	Provides concrete copy-paste blocks (hook JSON, mkdir commands, `python3 instinct-cli.py promote --dry-run`), but the repeatedly-referenced `observe.sh` is not present in the bundle, so a core executable piece is missing.	2 / 3
Workflow Clarity	The observe→detect→create→evolve sequence is shown in the diagram and promote has a --dry-run checkpoint, but the broader automated learning workflow lacks explicit validation gates despite involving batch/auto operations.	2 / 3
Progressive Disclosure	The body is well-sectioned but monolithic — reference-like content (config tables, file structure, scope guide, version diffs) is all inline with no split into reference files, and the referenced observe.sh is absent from the bundle.	2 / 3
	Total	8 / 12 Passed

Description

60%

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 specific and clearly conveys the system's concrete actions in third person, but it is jargon-heavy and lacks any explicit "Use when..." trigger clause, leaving the activation conditions implicit.

Suggestions

Append a "Use when..." clause with natural trigger terms (e.g., "Use when setting up automatic session learning, managing learned behaviors, or evolving repeated patterns into skills/commands/agents").

Soften the jargon ("原子本能"/"置信度评分") toward language a user would actually say, and add common variations like "learn from sessions", "remember patterns", "reuse knowledge".

Add a distinguishing phrase that separates this from general memory/learning skills, e.g. tying activation to hook-based observation and instinct evolution.

Dimension	Reasoning	Score
Specificity	Lists several concrete actions — "通过钩子观察会话", "创建带有置信度评分的原子本能", "演化为技能/命令/智能体" — matching the anchor for multiple specific concrete actions, in third-person voice.	3 / 3
Completeness	Clearly answers what the system does but provides no explicit "Use when..." trigger guidance, so per the rubric a missing trigger clause caps completeness at 2.	2 / 3
Trigger Term Quality	Contains some relevant/natural terms (学习系统, 技能, 命令, 智能体) but they are wrapped in heavy jargon (原子本能, 置信度评分, 项目作用域) and miss the natural user phrasing like "learn from my sessions" or "remember my patterns".	2 / 3
Distinctiveness Conflict Risk	The instinct/hook-learning niche is fairly specific, but the absence of explicit triggers and the broad learning/memory framing means it could still overlap with other learning or memory skills.	2 / 3
	Total	9 / 12 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: xu-xiang/everything-claude-code-zh
Path: docs/ja-JP/skills/continuous-learning-v2/SKILL.md
Commit: dfbf946

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.