Content
77%Weight 40%Scale 1-3Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
A well-structured, actionable legal-review skill with a clear sequenced workflow and strong validation gates. Its weaknesses are token efficiency (tutor-style explanations of at-will concepts and repeated boilerplate) and a monolithic single-file layout that could move jurisdiction detail into a reference.
Suggestions
Tighten the at-will section: drop the general definition ('At-will means either party can terminate...') and condense the seven-jurisdiction primer into a short table, keeping only the non-obvious statutory specifics (Montana WDEA, UK ERA 1996 s.1).
Factor the repeated '~/.claude/plugins/config/claude-for-legal/employment-legal/CLAUDE.md' path into a single defined alias used throughout, and consolidate the recurring 'Cite primary sources. Verify currency.' into one stated rule instead of repeating it per step.
Move the jurisdiction-by-jurisdiction at-will / written-statement particulars into a references file (e.g. references/at-will-by-jurisdiction.md) and link to it one level deep, so SKILL.md stays an overview.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is mostly efficient and adds genuinely specialized material (Montana WDEA, UK ERA 1996 s.1, the no-silent-supplement policy, source-attribution tags), but it also explains concepts Claude already knows — e.g. "'At-will' means either party can terminate without cause or notice" and a seven-jurisdiction at-will primer — and repeats 'Cite primary sources. Verify currency.' several times, so it sits at the 'mostly efficient but includes some unnecessary explanation' anchor rather than the lean top anchor. | 2 / 3 |
Actionability | Guidance is concrete and copy-ready: a numbered 5-step workflow, a classification test table, an offer-letter content checklist, exact escalation phrasings ('First hire in [state]... Research needed before offer goes out'), a markdown output template, and explicit source-attribution tags ([Westlaw], [CourtListener], [web search — verify]). Per the rubric's code-vs-instruction note, the absence of executable code is not penalized because the instruction-level guidance is highly actionable. | 3 / 3 |
Workflow Clarity | Steps are clearly sequenced (Jurisdiction → Classification → Restrictive covenants → Jurisdiction-specific → Offer letter) with explicit validation checkpoints: 'Research before calling exemption', 'Research enforceability before advising', the 'No silent supplement' report-and-stop loop, and the 'Consequential-action gate' that blocks a 'Clear to send' output without an explicit yes — matching the 'clear sequence with explicit validation steps and feedback loops' anchor. | 3 / 3 |
Progressive Disclosure | No bundle files exist (references/, scripts/, assets/ are absent), and while the file is organized into clear sections and delegates some detail to a one-level-deep external config path (~/.claude/plugins/.../CLAUDE.md), that single reference is repeated inline many times and substantive material like the seven-jurisdiction at-will primer lives inline rather than in a separate reference — so it is 'some structure, could be better organized' rather than the clean multi-file top anchor. | 2 / 3 |
Total | 10 / 12 Passed |