Content
53%Weight 40%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The body is well-organized and rich in genuinely non-obvious startup-GC domain knowledge with clearly sequenced workflows, but the skill as shipped is hollow: all five files it instructs the reader to run or consult (two scripts, three references) are missing from the bundle, and the workflows lack validation checkpoints. The SKILL.md works as an overview of a bundle that does not exist.
Suggestions
Ship the referenced bundle — add scripts/contract_risk_scanner.py, scripts/term_sheet_analyzer.py, and references/contracts_playbook.md, references/ip_and_regulatory.md, and references/term_sheet_decoder.md — or remove the references and inline the essential checklists so no instruction points at a nonexistent file.
Add validation checkpoints to the workflows (e.g., after the risk scan, confirm every HIGH finding is addressed by a counter-proposal before proceeding; in the IP audit, confirm the OSS inventory actually completed before flagging copyleft violations).
Trim the '## Keywords' section and the Version/Status/Disclaimer footer, which duplicate the frontmatter description, trigger terms, and metadata.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Mostly dense, non-obvious domain guidance Claude does not already have ("1x non-participating is standard; 1x participating or 2x is hostile"; "5-year auto-renew with 60-day notice is a common founder mistake") with no padding explaining known concepts. Not 5 because the ~40-item "## Keywords" section and the "Version / Status / Disclaimer" footer duplicate the frontmatter description, triggers, and metadata, and the version/status footer is time-sensitive boilerplate outside any deprecated section. | 4 / 5 |
Actionability | Concrete commands are specified ("python scripts/contract_risk_scanner.py path/to/contract.txt", "Run term_sheet_analyzer.py to get a 0-100 founder-friendliness score"), but neither script exists in the bundle — references/, scripts/, and assets/ are absent entirely — and the detailed checklists and glossaries are deferred to three references/*.md files that are also missing. Not 4 because missing tooling and missing referenced detail are more than minor gaps; not 2 because the body itself still supplies specific steps, negotiation standards, and a regulatory trigger table. | 3 / 5 |
Workflow Clarity | Four workflows are clearly numbered and sequenced ("1. Save the contract as plain text 2. Run contract_risk_scanner.py... 3. For each HIGH risk finding, draft a counter-proposal"), but no validation or verification checkpoints appear anywhere, and the batch IP Hygiene Audit ("Confirm every employee and contractor (past 12 months) signed invention assignment") has no verification step — the batch-operation cap of 3 applies. Steps also depend on scripts and "/cs:decide" / "/cs:freeze" commands not present in this bundle. | 3 / 5 |
Progressive Disclosure | Sections are well organized and references are clearly signaled ("See references/ip_and_regulatory.md"; a dedicated References section with one-line descriptions), but scored against the actual bundle structure, none of the five referenced files exist — both scripts and all three reference docs are absent, so every path to deeper material dead-ends like the bad "See advanced.md for more details" example. The disclosure structure is non-functional as shipped; not 3 because the failure is not mere organization but a completely empty referenced layer. | 2 / 5 |
Total | 12 / 20 Passed |