yes-md

6-layer AI governance: safety gates, evidence-based debugging, anti-slack detection, and machine-enforced hooks. Makes AI safe, thorough, and honest.

Quality

40%

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 ./skills/yes-md/SKILL.md

Quality

Content

52%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 skill provides a comprehensive AI governance framework with genuinely useful structures like safety gates, debugging escalation levels, and anti-slack detection. However, it is significantly over-long due to redundant content (the Evidence Table repeats earlier sections), philosophical framing that doesn't add actionable value, and explanations of behaviors Claude already understands. The monolithic structure with no progressive disclosure makes it token-expensive to load.

Suggestions

Remove the Evidence Table section entirely — it duplicates the Anti-Slack Detection table and the Three Iron Rules with no new information.

Cut the opening philosophical framing (paragraphs 2-4 about PUA vs YES, encouragement vs intimidation) to a single sentence; Claude doesn't need motivational context.

Split into multiple files: keep the Three Iron Rules and Safety Gates in SKILL.md, move Debugging Escalation to DEBUGGING.md, Ripple Check + Bug Closure to VERIFICATION.md, and Anti-Slack Detection to ANTI-SLACK.md with clear references.

Remove the 'Seven Deadly Shortcuts' table — the Anti-Slack Detection table already covers the same behaviors with self-correction actions, making the earlier table redundant.

Dimension	Reasoning	Score
Conciseness	Significantly verbose. The opening philosophical framing ('PUA says NO. YES says YES', encouragement vs intimidation), the 'Seven Deadly Shortcuts' table, the 'Evidence Table' at the end (which largely repeats earlier content), and the compatibility section all add substantial token overhead. Many concepts explained here (like 'verify your changes' or 'don't guess') are things Claude already knows — the skill over-explains the rationale rather than just stating the rules.	2 / 5
Actionability	Provides concrete commands (cp, curl, grep, ls -la, lsof, node -v), specific banned phrases, and clear do/don't examples with bash snippets. However, much of the guidance is still meta-instruction ('run the verification command first') rather than fully executable code blocks. The checklists are actionable but some items remain somewhat abstract (e.g., 'handle concurrent access').	4 / 5
Workflow Clarity	The debugging escalation levels (2-5+ failures) provide a clear, well-sequenced workflow with explicit mandatory actions at each level. Safety gates are well-structured with triggers and actions. The Bug Closure Protocol and Ripple Check provide validation steps. Minor gap: the gates lack explicit feedback loops (e.g., what to do if backup fails, or if blast radius check reveals issues beyond simple 'investigate before changing').	4 / 5
Progressive Disclosure	This is a monolithic ~200+ line file with no references to external files and no bundle structure. The Evidence Table at the end largely duplicates the Anti-Slack Detection table and earlier rules. Content like the debugging escalation protocol, safety gates, and ripple checks could be split into separate referenced files. Everything is inlined in one long document.	2 / 5
	Total	12 / 20 Passed

Description

28%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.

This description relies heavily on buzzwords and marketing-style language ('6-layer AI governance', 'Makes AI safe, thorough, and honest') without explaining concrete actions or when the skill should be selected. It lacks a 'Use when...' clause and uses jargon that users are unlikely to naturally invoke, making it difficult for Claude to correctly select this skill from a pool of alternatives.

Suggestions

Add an explicit 'Use when...' clause with natural trigger terms, e.g., 'Use when setting up code review pipelines, enforcing safety checks, or adding pre-commit hooks for AI-generated code.'

Replace buzzwords with concrete actions using third-person verbs, e.g., 'Enforces safety gates on AI outputs, runs evidence-based debugging checks, detects shortcut-taking in AI responses, and configures git hooks for automated validation.'

Include natural keywords users would actually say, such as 'code quality', 'safety checks', 'pre-commit hooks', 'AI output validation', or 'automated review'.

Dimension	Reasoning	Score
Specificity	Names some concepts like 'safety gates', 'evidence-based debugging', 'anti-slack detection', and 'machine-enforced hooks', but these are more like buzzwords than concrete actions. No specific verbs describe what the skill actually does (e.g., 'validates', 'enforces', 'audits').	2 / 5
Completeness	The 'what' is vaguely described through buzzwords ('6-layer AI governance') but lacks concrete actions. There is no 'when' clause or trigger guidance at all — no 'Use when...' or equivalent. The final sentence 'Makes AI safe, thorough, and honest' is a marketing claim, not a usage trigger.	2 / 5
Trigger Term Quality	Terms like 'AI governance', 'safety gates', and 'anti-slack detection' are specialized jargon unlikely to match natural user queries. Users would more likely say things like 'code review', 'safety checks', or 'quality assurance' rather than these coined terms.	2 / 5
Distinctiveness Conflict Risk	The niche of 'AI governance' with specific sub-concepts like 'anti-slack detection' and 'machine-enforced hooks' is somewhat distinctive, but the vagueness of the description and broad claims like 'Makes AI safe, thorough, and honest' could cause confusion with other quality assurance or code review skills.	3 / 5
	Total	9 / 20 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: administrakt0r/AI-Agents-Safe-Coding-Skills
Path: skills/yes-md/SKILL.md
Commit: 900c504

Reviewed: 2 hours 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.