ai-md

Convert human-written CLAUDE.md into AI-native structured-label format. Battle-tested across 4 models. Same rules, fewer tokens, higher compliance.

Quality

61%

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/AI-Agents-Safe-Coding-Skills-claude/skills/ai-md/SKILL.md

The canonical home for this skill is ai-md in sickn33/agentic-awesome-skills

Quality

Content

70%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.

The body delivers a highly actionable, well-sequenced methodology with strong validation feedback loops for a destructive operation. Its main weakness is conciseness and the absence of progressive disclosure — a 510-line monolith that should offload the template, label table, and test protocol into reference files.

Suggestions

Move the AI-native template, the label-vocabulary table, and the 8-question exam protocol into separate reference files (e.g. references/template.md, references/labels.md, references/exam-protocol.md) and link them one level deep from SKILL.md.

Cut marketing/explanatory prose such as 'The paradox we proved', 'The uncomfortable truth', and narrative mechanism explanations that restate what Claude already knows about attention.

Tighten the 'Why It Works' section to a brief rationale plus a single example rather than three multi-paragraph mechanisms.

Dimension	Reasoning	Score
Conciseness	The body is mostly efficient with concrete examples (label tables, phase breakdowns, anti-patterns), but retains explanatory and marketing prose ('LLMs don't read — they attend', 'The uncomfortable truth...') that could be trimmed, fitting 'mostly efficient but includes some unnecessary explanation'.	3 / 5
Actionability	It provides a concrete six-phase process, a full label-vocabulary table, a copy-paste bash measurement script, the AI-native template, and an 8-question exam protocol — mostly executable guidance with only minor gaps, sitting above the pseudocode anchor but short of fully copy-paste-ready for the conversion itself.	4 / 5
Workflow Clarity	The two-stage PREVIEW/DISTILL workflow is explicitly sequenced with a backup step before the destructive CLAUDE.md overwrite, a multi-model validation gate (Phase 6), and a revert-on-regression feedback loop, matching the 'clear sequence with explicit validation and feedback loops' anchor.	5 / 5
Progressive Disclosure	The document is well-headered and navigable, but at ~510 lines it keeps the template, label-vocabulary table, and test protocol all inline rather than splitting them into one-level-deep reference files, fitting 'some structure but content that should be separate is inline'.	3 / 5
	Total	15 / 20 Passed

Description

53%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.

The description states a clear, distinctive purpose but omits explicit trigger guidance and relies on marketing language rather than natural user keywords. Adding a concrete 'Use when...' clause with the body's existing triggers would lift completeness and trigger-term quality.

Suggestions

Add an explicit 'Use when...' clause naming natural user triggers such as 'your CLAUDE.md is long but AI still ignores your rules' or 'token usage is too high from verbose system instructions'.

Replace the marketing tail ('Battle-tested across 4 models. Same rules, fewer tokens, higher compliance.') with 1-2 additional concrete actions or synonyms (e.g. 'migrating rules between AI tools').

Include common synonyms/extensions (CLAUDE.md, system instructions, system prompt) so natural user phrasing matches the skill.

Dimension	Reasoning	Score
Specificity	The description names the domain ('CLAUDE.md') and one concrete action ('Convert ... into AI-native structured-label format'), but stops at a single action without listing the constituent conversion steps, matching the '1-2 concrete actions, not comprehensive' anchor.	3 / 5
Completeness	A clear 'what' is present (convert CLAUDE.md to structured-label format) but there is no 'Use when...' clause or equivalent explicit trigger guidance, so completeness is capped at 3 per the guideline.	3 / 5
Trigger Term Quality	It surfaces relevant keywords ('CLAUDE.md', 'tokens', 'models') but leans on marketing phrasing ('Battle-tested', 'higher compliance') and omits the natural trigger phrases a user would actually say (e.g. 'AI ignores my rules', 'system prompt too long'), matching the 'some relevant keywords but missing common variations' anchor.	3 / 5
Distinctiveness Conflict Risk	The CLAUDE.md-to-structured-label conversion niche is specific and clearly distinguishable, with only minor overlap risk against general prompt-engineering skills, fitting the 'mostly distinct; minor overlap' anchor better than the broader 3 or the uniquely-niche 5.	4 / 5
	Total	13 / 20 Passed

Validation

87%

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 — 14 / 16 Passed

Validation for skill structure

Criteria	Description	Result
skill_md_line_count	SKILL.md is long (519 lines); consider splitting into references/ and linking	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	14 / 16 Passed

Repository: administrakt0r/AI-Agents-Safe-Coding-Skills
Path: plugins/AI-Agents-Safe-Coding-Skills-claude/skills/ai-md/SKILL.md
Commit: 900c504

Reviewed: 14 minutes 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.