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

58%

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%Weight 40%Scale 1-3

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-structured and offers concrete setup/config guidance, but it is padded with redundancy, leaves the core observer/evolution logic as diagrams rather than executable code, and lacks validation checkpoints plus any reference-file split despite its length.

Suggestions

De-duplicate content: keep commands and confidence scoring in one place, and collapse the plugin/manual hook snippets into a single example with a one-line variant note.

Add a validation/feedback loop to the evolution workflow (e.g. review generated skills against the source instincts and re-cluster on rejection) so generative batch operations have explicit checkpoints.

Move the full config.json reference, the confidence-scoring tables, and the architecture diagram into reference files under references/ and link to them from SKILL.md to apply progressive disclosure.

Dimension	Reasoning	Score
Conciseness	Mostly efficient with concrete config and examples, but includes redundancy (commands listed in both 快速入门 and the 命令 table; confidence scoring repeated across the instinct model and a separate section; dual plugin/manual hook snippets) plus v1/v2 marketing and a justification section that could be trimmed.	2 / 3
Actionability	Provides copy-paste-ready hook JSON, bash init commands, a config.json template, and an instinct YAML, but the core logic (observe.sh body, observer agent pattern detection, /evolve clustering) is shown only as an ASCII diagram, not as executable code.	2 / 3
Workflow Clarity	Quick start and the observation→instinct→evolution pipeline are sequenced, but there are no validation/verification checkpoints for the generative batch operation of evolving instincts into skills/commands/agents, which caps workflow clarity at 2.	2 / 3
Progressive Disclosure	No bundle files exist (references/, scripts/, assets/ absent) and the ~280-line body keeps config reference, confidence-scoring detail, and architecture inline rather than splitting them into one-level-deep reference files; sections are organized but the content is monolithic.	2 / 3
	Total	8 / 12 Passed

Description

67%Weight 40%Scale 1-3

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 concretely names several capabilities and occupies a distinctive niche, but it omits any explicit "Use when..." trigger guidance and relies on jargon-heavy terms rather than natural user phrasing.

Suggestions

Append an explicit trigger clause, e.g. "Use when you want Claude to learn recurring patterns from your sessions or evolve learned behaviors into reusable skills/commands/agents."

Soften jargon by adding natural-language variants users would actually say (e.g. "learning from my sessions", "remembering my workflows", "building habits") alongside the Instinct/Hook terminology.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions — "通过钩子观察会话" (observe sessions via hooks), "创建带有置信度评分的原子化本能" (create confidence-scored atomic instincts), "进化为技能、命令或智能体" (evolve into skills/commands/agents) — matching the multi-action anchor.	3 / 3
Completeness	Clearly states what the system does but provides no "Use when..." or equivalent explicit trigger guidance, which per the guidelines caps completeness at 2.	2 / 3
Trigger Term Quality	Contains some relevant domain terms (训练系统/training system, 观察/observe, 本能/instinct) but leans heavily on technical jargon ("Hooks", "置信度评分", "原子化本能") a user would not naturally say, and omits common variations.	2 / 3
Distinctiveness Conflict Risk	The instinct-based learning niche with hooks, confidence scoring, and evolution into skills/commands/agents is a distinct, specific domain unlikely to conflict with typical task skills.	3 / 3
	Total	10 / 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: 28 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.