cookbook-audit

Audit an Anthropic Cookbook notebook based on a rubric. Use whenever a notebook review or audit is requested.

1.96x

Quality

63%

Does it follow best practices?

Impact

100%

1.96x

Average score across 3 eval scenarios

Securityby

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./.claude/skills/cookbook-audit/SKILL.md

Quality

Content

56%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 skill is well-structured and gives concrete, actionable audit guidance with a clear workflow and report template. Its main weaknesses are heavy duplication of style_guide.md content inside SKILL.md and missing bundle files the body tells Claude to read first.

Suggestions

Move the "Style Guidelines", "Structural Requirements", "Content Philosophy", "What Cookbooks Are NOT", and "Common Anti-Patterns" sections into style_guide.md, leaving SKILL.md as a concise overview that points to it.

Remove general writing best-practice bullets (active voice, short paragraphs, use headers) that Claude already applies; keep only Cookbook-specific conventions.

Ensure the referenced style_guide.md and validate_notebook.py actually ship in the skill bundle so the "Always read style_guide.md first" instruction is satisfiable.

Dimension	Reasoning	Score
Conciseness	The 273-line body noticeably pads with sections that duplicate the referenced style_guide.md ("Style Guidelines", "Structural Requirements", "What Cookbooks Are NOT", "Common Anti-Patterns") and restate general writing advice Claude already knows ("Active voice preferred", "Short paragraphs (3-5 sentences)").	2 / 5
Actionability	Concrete guidance is present — an executable command ("python3 validate_notebook.py <path>"), specific bundle paths ("scripts/detect-secrets/plugins.py", ".secrets.baseline", "tmp/"), a filled-in audit report template, and named scoring dimensions — with only minor gaps because the scoring anchors live in the external style_guide.md.	4 / 5
Workflow Clarity	An explicit 8-step numbered workflow plus a Quick Reference Checklist gives a clear sequence, and detect-secrets runs as automated validation in step 3, but there is no explicit feedback loop telling Claude what to do when the checks surface issues.	4 / 5
Progressive Disclosure	Headers and signaled references to style_guide.md provide structure, yet roughly 140 lines of style/structural/anti-pattern guidance that the skill itself says the style guide canonically holds are inlined into SKILL.md, and the referenced style_guide.md and validate_notebook.py are absent from the bundle.	3 / 5
	Total	13 / 20 Passed

Description

70%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 clearly states what the skill does and when to invoke it, with a concrete trigger clause and a fairly distinct niche. It is held back from the top band by a single named action and a thin set of trigger synonyms.

Suggestions

Broaden the trigger clause to additional natural phrases, e.g. "Use whenever a notebook review, audit, scoring, or quality check is requested."

Add a second concrete action to the 'what' portion, e.g. "Audit ... and generate a scored report with improvement recommendations."

Dimension	Reasoning	Score
Specificity	Quotes "Audit an Anthropic Cookbook notebook based on a rubric" name the domain and one concrete action (audit) but no second concrete action is stated, matching the anchor for 1-2 concrete actions without comprehensive coverage.	3 / 5
Completeness	Both halves are present — a clear "what" ("Audit an Anthropic Cookbook notebook based on a rubric") and an explicit "when" ("Use whenever a notebook review or audit is requested"), but the trigger clause lists only two variations rather than the broader, more specific set that would merit a 5.	4 / 5
Trigger Term Quality	Natural keywords appear ("notebook", "cookbook", "review", "audit", "rubric") that a user would plausibly say, giving good coverage, though a few common variations (grade, score, evaluate, check) are missing.	4 / 5
Distinctiveness Conflict Risk	The niche is specific (Anthropic Cookbook notebooks, audit/review), keeping it mostly distinct with only minor overlap risk against a generic code-review or notebook-review skill.	4 / 5
	Total	15 / 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
referenced_paths_exist	Referenced path issues: 2 missing, 2 deeper-than-1-level	Warning

	Total	15 / 16 Passed

Repository: anthropics/claude-cookbooks
Path: .claude/skills/cookbook-audit/SKILL.md
Commit: 35f2eec

Reviewed: 4 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.