continuous-learning

Automatically extract reusable patterns from Claude Code sessions and save them as learned skills for future use.

Quality

—

Does it follow best practices?

Impact

—

No eval scenarios have been run

Securityby

Passed

No known issues

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-structured and gives concrete config and hook-setup snippets, but it leans on a referenced evaluate-session.sh that is not bundled and pads the document with a dated competitor-comparison section. Actionability and workflow clarity suffer from the missing execution logic and absent validation steps. Moving the research/comparison material to a real referenced file would improve both conciseness and progressive disclosure.

Suggestions

Provide the actual evaluate-session.sh (in a scripts/ bundle) or inline its core logic so the session-evaluation, pattern-detection, and skill-extraction steps are executable rather than just described.

Add explicit validation/verification checkpoints, e.g. confirm a detected pattern meets the extraction_threshold and verify a written skill file is valid before finalizing.

Move the "比較筆記" Homunculus comparison and v2 roadmap into a referenced COMPARISON.md or docs/ file (and create it) so the dated, non-essential material is one level deep instead of inline.

Dimension	Reasoning	Score
Conciseness	The core setup is efficient, but the dated "比較筆記（研究：2025 年 1 月）" Homunculus comparison and v2 roadmap are non-essential padding; the time-stamped research section also penalizes conciseness per the guidelines.	2 / 3
Actionability	config.json and the hook-setup JSON are concrete and copy-paste ready, but the core evaluate-session.sh logic is referenced yet not provided in any scripts/ bundle, leaving the main task described rather than executable.	2 / 3
Workflow Clarity	The three steps (evaluate, detect, extract) are sequenced, but there are no validation/verification checkpoints for a skill that writes files into ~/.claude/skills/learned/, which caps workflow clarity at 2.	2 / 3
Progressive Disclosure	Sections are organized, but the body references a non-existent docs/continuous-learning-v2-spec.md and no references/scripts/assets bundle is present; the inline Homunculus comparison is content that should live in a separate file.	2 / 3
	Total	8 / 12 Passed

Description

50%

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 clear and grammatical in third person, stating what the skill does, but it omits any explicit "Use when..." trigger guidance. This leaves completeness and trigger discoverability at the mid level. Adding natural-language trigger conditions would lift the strongest weighted dimensions.

Suggestions

Append a 'Use when...' clause naming natural triggers, e.g. 'Use when ending a Claude Code session that surfaced reusable error fixes, workarounds, or debugging techniques worth saving as a skill.'

Expand the action list beyond extract/save (e.g. 'detect, extract, deduplicate, and save') to reach the level-3 specificity bar.

Add common user phrasings like 'session patterns', 'learned skills', 'remember this fix' so the description surfaces for natural requests.

Dimension	Reasoning	Score
Specificity	It names the domain ("Claude Code sessions") and two concrete actions ("extract reusable patterns", "save them as learned skills"), but only two actions rather than the multiple specific actions of a level-3 example.	2 / 3
Completeness	It clearly states what the skill does, but the "when" is only implied and there is no explicit "Use when..." clause, which per the guidelines caps completeness at 2.	2 / 3
Trigger Term Quality	"Claude Code sessions" and "learned skills" are relevant keywords a user might say, but coverage of natural trigger variations is thin and there is no "Use when..." phrasing.	2 / 3
Distinctiveness Conflict Risk	The Claude-Code-session/learned-skills niche is fairly specific, but with no explicit triggers it could still overlap with other learning or meta-skills.	2 / 3
	Total	8 / 12 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: ysyecust/everything-claude-code
Commit: a9cbecb

Reviewed: 1 day 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.