ai-tool-handoff

Detects when Luminance, Kira, or a similar bulk-review tool is in use, hands off the high-volume clause extraction to it, and QAs its output per the trust level in `~/.claude/plugins/config/claude-for-legal/corporate-legal/CLAUDE.md`. Use when user says "send to Luminance", "bulk review", "AI extraction", or when diligence-issue-extraction hits a high-volume category.

Quality

92%

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

85%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 well-structured, actionable skill body with a clear multi-step workflow and an explicit QA feedback loop. Its main weakness is token efficiency: a very long config path is repeated many times and a few sentences editorialize rather than instruct.

Suggestions

Define the long config path once (e.g., 'the corporate-legal CLAUDE.md at ~/.claude/plugins/config/claude-for-legal/corporate-legal/CLAUDE.md (hereafter "the CLAUDE.md")') and reuse the short form throughout to cut repeated tokens.

Tighten the Purpose paragraph — 'Luminance and Kira are good at one thing...' restates the skill's value rather than instructing an action; lead with the decision guidance instead.

Consider moving the load-request and results-table markdown templates into a references/ file to keep SKILL.md as a lean overview, which would also let progressive disclosure show actual bundle structure.

Dimension	Reasoning	Score
Conciseness	Mostly efficient and it correctly assumes Claude's competence (no explanation of what a change-of-control clause or Luminance is), but the long path '~/.claude/plugins/config/claude-for-legal/corporate-legal/CLAUDE.md' is repeated ~8 times verbatim and the 'Luminance and Kira are good at one thing...' Purpose paragraph is mild padding — it could be tightened by defining the path once and trimming the editorializing.	2 / 3
Actionability	Gives concrete, copy-paste-ready guidance: a quantified '>50 documents' threshold, a specific clause-type list, a fill-in load-request template, and a results-table template — fully actionable rather than abstract.	3 / 3
Workflow Clarity	A clearly sequenced four-step process (Prepare batch → Load → QA → Judgment layer) with an explicit validation/feedback loop in the QA step ('If error rate is low, accept the batch. If errors found, widen the sample'), satisfying the checkpoint requirement for batch operations.	3 / 3
Progressive Disclosure	No bundle files are present, so all content lives in a single well-sectioned SKILL.md; per the rubric's simple-skills note, clear organization without external references earns the top score, and there are no deep or broken references.	3 / 3
	Total	11 / 12 Passed

Description

100%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 strong, third-person description that names concrete actions, gives natural user-facing trigger phrases, and explicitly covers both what and when. It is specific to a niche and unlikely to collide with sibling skills.

Dimension	Reasoning	Score
Specificity	Lists several concrete actions — 'Detects when Luminance, Kira... is in use', 'hands off the high-volume clause extraction to it', and 'QAs its output per the trust level' — matching the multiple-specific-actions anchor.	3 / 3
Completeness	Explicitly answers both what (detect, hand off, QA) and when via a clear 'Use when user says...' clause with multiple triggers, matching the top anchor.	3 / 3
Trigger Term Quality	Includes natural quoted phrases a user would actually say — 'send to Luminance', 'bulk review', 'AI extraction' — giving good coverage of natural trigger terms.	3 / 3
Distinctiveness Conflict Risk	Occupies a clear niche (bulk AI-tool handoff + QA) with distinct triggers and an explicit contrast against the sibling diligence-issue-extraction skill, making wrongful triggering unlikely.	3 / 3
	Total	12 / 12 Passed

Validation

100%

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 — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: anthropics/claude-for-legal
Path: corporate-legal/skills/ai-tool-handoff/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.