continuous-learning-v2

Instinct-based learning system that observes sessions via hooks, creates atomic instincts with confidence scoring, and evolves them into skills/commands/agents.

1.58x

Quality

26%

Does it follow best practices?

Impact

89%

1.58x

Average score across 6 eval scenarios

Securityby

Advisory

Suggest reviewing before use

Optimize this skill with Tessl

npx tessl skill review --optimize ./docs/zh-TW/skills/continuous-learning-v2/SKILL.md

Quality

Content

35%

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

This skill provides a comprehensive conceptual overview of an instinct-based learning system but suffers from significant verbosity and incomplete actionability. The architecture is well-diagrammed and setup steps are concrete, but critical executable components (hook scripts, observer agent, slash command implementations) are referenced without being defined. The document reads more like a product README than an efficient skill instruction file.

Suggestions

Remove or drastically shorten conceptual sections (v1 vs v2 comparison, 'Why Hooks vs Skills', backward compatibility, privacy) — Claude doesn't need persuasion, just instructions.

Include the actual content of observe.sh hook script and slash command implementations, or clearly indicate they must be created and provide their expected behavior/interface.

Add validation checkpoints: after hook setup, verify with a test command that observations.jsonl is receiving data; after observer start, confirm it's running.

Move the full config.json, confidence scoring table, and file structure details into separate reference files, keeping only a minimal quick-start config inline.

Dimension	Reasoning	Score
Conciseness	The skill is extremely verbose at ~200+ lines. It includes extensive explanations of concepts (what instincts are, why hooks vs skills, backward compatibility), a comparison table with v1, detailed architecture diagrams, and confidence scoring explanations that Claude doesn't need explained at this length. Much of this is conceptual padding rather than actionable instruction.	1 / 3
Actionability	It provides concrete JSON config, bash commands for setup, and directory structure creation commands that are copy-paste ready. However, the actual hook script (observe.sh), observer agent script (start-observer.sh), and the slash commands are referenced but never defined or shown — the core executable pieces are missing, making this only partially actionable.	2 / 3
Workflow Clarity	The quick start section provides a numbered sequence (enable hooks → init directories → run observer), and the architecture diagram shows the data flow clearly. However, there are no validation checkpoints — no way to verify hooks are working, no way to confirm observations are being recorded, and no error recovery steps if something fails.	2 / 3
Progressive Disclosure	The content has clear sections and references external files (config.json, various scripts, external links). However, too much detail is inline (full config.json, confidence scoring tables, backward compatibility notes, privacy section) that could be in separate reference files. The external references at the bottom (Skill Creator, Longform Guide) are appropriate but the main body is bloated.	2 / 3
	Total	7 / 12 Passed

Description

17%

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

This description reads like an internal architecture summary rather than a functional skill description. It relies heavily on custom jargon ('atomic instincts', 'confidence scoring', 'hooks') that would not help Claude match user requests to this skill. It lacks both natural trigger terms and explicit 'when to use' guidance.

Suggestions

Add a 'Use when...' clause that describes concrete scenarios, e.g., 'Use when the user wants to automatically learn patterns from sessions and generate reusable skills or commands.'

Replace jargon with natural language trigger terms users might say, such as 'learn from usage', 'auto-generate skills', 'session patterns', 'adaptive learning'.

Describe concrete user-facing actions instead of internal mechanisms, e.g., 'Monitors coding sessions to identify repeated patterns, then generates reusable skills, commands, or agents based on observed behavior.'

Dimension	Reasoning	Score
Specificity	Names some actions like 'observes sessions via hooks', 'creates atomic instincts with confidence scoring', and 'evolves them into skills/commands/agents', but these are more conceptual/architectural descriptions than concrete user-facing actions. The language is domain-specific jargon rather than clear capabilities.	2 / 3
Completeness	Provides a vague 'what' using abstract terminology but completely lacks any 'when should Claude use it' guidance. No 'Use when...' clause or equivalent trigger guidance, which per the rubric should cap completeness at 2, but the 'what' is also weak, so this scores a 1.	1 / 3
Trigger Term Quality	Uses highly technical jargon like 'atomic instincts', 'confidence scoring', 'hooks' that users would almost never naturally say. No natural trigger terms a user would use when needing this functionality.	1 / 3
Distinctiveness Conflict Risk	The concept of 'instinct-based learning' is somewhat unique and unlikely to directly conflict with common skills, but the mention of 'skills/commands/agents' is broad enough to potentially overlap with other meta-skills or automation tools.	2 / 3
	Total	6 / 12 Passed

Validation

90%

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 — 10 / 11 Passed

Validation for skill structure

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

	Total	10 / 11 Passed

Repository: ysyecust/everything-claude-code
Commit: 79cc4e3

Reviewed: 3 months 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.