vendor-ai-review

Review vendor AI terms — agreement, addendum, or ToS AI provisions — against your governance positions; flag training-on-data, liability, model changes, and AI policy consistency. Use when user says "review this AI agreement", "check OpenAI terms", "what did we agree to with [vendor]", "vendor sent an AI addendum", "is this AI contract okay", or attaches vendor AI terms.

Quality

83%

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

Quality

Content

77%Weight 40%Scale 1-3

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A strong, highly actionable legal-workflow skill with clear sequencing and validation gates. Its main weaknesses are a verbose explanatory tail and a monolithic structure that, given its length, would benefit from splitting detailed reference material into bundled files.

Suggestions

Trim the "Practical notes" and "Purpose" narrative prose to only the non-obvious, skill-specific guidance Claude would not already infer.

Move the long output template and/or the term-by-term table into a bundled reference file (e.g. references/output-template.md) and link to it from the body, improving progressive disclosure for a skill of this length.

Reconcile the top 7-step summary with the detailed body sections to remove overlap and shorten the document.

Dimension	Reasoning	Score
Conciseness	Mostly task-specific and actionable, but the ~310-line body includes explanatory prose (e.g. the "Practical notes" section's narrative on training-on-data history and the Purpose section's framing) that could be tightened without losing guidance value.	2 / 3
Actionability	Highly executable: a 13-row term checklist with per-term "what to look for", copy-paste output templates, ready redline language (e.g. the flow-down clause text), explicit severity ratings, and a concrete invocation example.	3 / 3
Workflow Clarity	Clear numbered sequence (steps 1-7) with explicit validation checkpoints — confirm governance positions populated, surface a [PLACEHOLDER] bounce, provisional-mode branch, and a hard signature gate requiring an explicit yes before proceeding.	3 / 3
Progressive Disclosure	Well-sectioned and it delegates detailed positions to an external one-level playbook file, but with no bundle files provided the ~310-line body is monolithic — the output template, term table, and practical notes are inline content that could be split into reference files.	2 / 3
	Total	10 / 12 Passed

Description

90%Weight 40%Scale 1-3

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 precise, well-triggered description that clearly conveys capability and firing conditions for a distinct niche. The only weakness is a second-person possessive ("your governance positions") that violates the third-person voice guideline and caps specificity at 2.

Suggestions

Rewrite in third person to avoid the voice penalty, e.g. "Reviews vendor AI terms ... against the user's governance positions" instead of "your governance positions".

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions ("Review vendor AI terms ... flag training-on-data, liability, model changes, and AI policy consistency"), which would anchor at 3, but the second-person phrase "against your governance positions" triggers the rubric's voice penalty, reducing specificity by one to 2.	2 / 3
Completeness	Explicitly answers both what (review/flag specific vendor AI term categories) and when (a named "Use when user says..." clause with concrete triggers), matching the score-3 anchor.	3 / 3
Trigger Term Quality	Strong coverage of natural phrases a user would say — "review this AI agreement", "check OpenAI terms", "what did we agree to with [vendor]", "vendor sent an AI addendum", "is this AI contract okay", plus "attaches vendor AI terms".	3 / 3
Distinctiveness Conflict Risk	Clear niche — vendor AI terms review against governance positions — with distinct, AI-specific triggers unlikely to fire for generic contract or DPA review skills.	3 / 3
	Total	11 / 12 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
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	15 / 16 Passed

Repository: anthropics/claude-for-legal
Path: ai-governance-legal/skills/vendor-ai-review/SKILL.md
Commit: 4a6c651

Reviewed: 6 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.