learn-and-improve

This skill should be used when the user asks to "improve my setup", "learn from this session", "fix my config", "stop asking for permissions", or reports friction with skills, agents, hooks, or permissions. Analyzes conversation history and proposes configuration improvements.

Quality

76%

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 ./plugins/toolkit/skills/learn-and-improve/SKILL.md

Quality

Content

78%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.

A well-structured, actionable six-step workflow with clean progressive disclosure to three real reference files and appropriate validation checkpoints. The chief gaps are minor: some inline tables duplicate the references, and the retry-on-invalid loop is implicit rather than stated.

Suggestions

Collapse the Step 3 root-cause and Step 4 reach tables to a few representative rows and point to improvement-types.md for the full catalog, reducing body length and overlap.

Make the validation feedback loop explicit in Step 5: 'If the file fails to parse, fix the syntax and re-verify before proceeding.'

Inline one complete settings.json permission-patch example so the body is copy-paste-ready without opening references.

Dimension	Reasoning	Score
Conciseness	Mostly efficient: uses tables, lists, and concrete file paths without explaining concepts Claude already knows. Below 5 because the inline root-cause table (Step 3) and reach table (Step 4) overlap material that also lives in references and could be trimmed.	4 / 5
Actionability	Provides concrete file paths, a permission-pattern example ('Bash(npm run *)'), symptom-to-fix tables, and a verified-improvement entry template. Below 5 because the fully executable JSON patches and decision tree are delegated to references rather than given copy-paste-ready in the body.	4 / 5
Workflow Clarity	Six steps are clearly sequenced with checkpoints (confirm-before-write, verify valid JSON/markdown, note restart needs). Validation is present so no cap applies, but below 5 because the body lacks an explicit 'if validation fails, fix and re-validate' retry loop.	4 / 5
Progressive Disclosure	The body is an overview that signals three real, one-level-deep references (improvement-types.md, verified-improvements.md, diagnostic-checklist.md) with clear link text and described purposes; all referenced files exist and content is appropriately split.	5 / 5
	Total	17 / 20 Passed

Description

73%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.

A strong, trigger-rich description that clearly signals when to use the skill and what it does, occupying a distinct meta-config niche. The main weakness is a somewhat thin 'what' that lists only two fairly general actions.

Suggestions

Expand the 'what' clause into 3-4 concrete actions (e.g., 'reads settings/CLAUDE.md/hooks/MCP config, pinpoints the root cause, and proposes scoped fixes') to lift specificity and completeness toward 5.

Add a couple of natural synonyms such as 'tune my setup' or 'make Claude stop prompting' to broaden trigger coverage.

Dimension	Reasoning	Score
Specificity	Names the domain and two concrete actions ('Analyzes conversation history' and 'proposes configuration improvements'), but the action list is not comprehensive. It is above 2 because the actions are specific rather than generic, and below 4 because there are only two rather than 'several specific actions'.	3 / 5
Completeness	Explicitly answers both 'when' (concrete trigger phrases via 'This skill should be used when...') and 'what' (analyzes conversation history, proposes configuration improvements). Below 5 because the 'what' portion is brief and less concrete than the multi-action anchor example.	4 / 5
Trigger Term Quality	Quotes several natural user phrases ('improve my setup', 'learn from this session', 'fix my config', 'stop asking for permissions') plus a friction surface (skills, agents, hooks, permissions). Good keyword coverage; below 5 because common synonyms/variations are missing.	4 / 5
Distinctiveness Conflict Risk	Occupies a clear meta-niche (diagnosing and improving Claude Code configuration friction) with distinctive triggers ('stop asking for permissions', friction with skills/agents/hooks) that are unlikely to fire for ordinary coding tasks, giving minimal conflict risk.	5 / 5
	Total	16 / 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
allowed_tools_field	'allowed-tools' contains unusual tool name(s)	Warning

	Total	15 / 16 Passed

Repository: dwmkerr/claude-toolkit
Path: plugins/toolkit/skills/learn-and-improve/SKILL.md
Commit: 4316684

Reviewed: 8 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.