Content
86%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-structured and highly actionable with real executable tooling and one-level-deep references, scoring high on actionability and progressive disclosure. The main weakness is workflow validation: sequences are clear but lack explicit checkpoint/feedback-loop steps.
Suggestions
Add an explicit validation checkpoint to Workflow 1 (e.g., 'Review HIGH-risk findings for false positives before drafting counter-proposals') and a sanity-check step after the scanner runs.
Insert a validate-fix-retry loop into Workflow 2 (e.g., re-run term_sheet_analyzer.py after negotiation changes to confirm the founder-friendliness score improved before logging).
Collapse the Key Questions section or cross-reference it to Core Responsibilities to remove the duplicated liability-cap and liquidation-preference content.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Dense and information-rich with no padding of concepts Claude already knows, but the Key Questions section duplicates content restated in Core Responsibilities (liability cap, liquidation preference), offering minor tightening opportunity. | 4 / 5 |
Actionability | Fully executable commands ('python scripts/contract_risk_scanner.py path/to/contract.txt'), concrete founder-friendly defaults ('1x non-participating is standard; 2x is hostile'), a trigger-to-regime-to-first-step table, and a copy-paste output template; the referenced scripts are real and stdlib-only. | 5 / 5 |
Workflow Clarity | Four clearly numbered workflows with an explicit 'before signing' gate and a /cs:freeze regret-lock, but validation checkpoints are largely implicit and there are no validate-fix-retry feedback loops. | 4 / 5 |
Progressive Disclosure | SKILL.md is a clear overview pointing to three real one-level-deep references (contracts_playbook.md, ip_and_regulatory.md, term_sheet_decoder.md) signaled both inline and in a References section, with bulk detail correctly pushed to the reference files. | 5 / 5 |
Total | 18 / 20 Passed |