Content
82%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 a tight, highly actionable GC checklist: six concrete questions with numeric thresholds, a copy-paste output template, executable script commands, and clear routing. The only notable weaknesses are one line of motivational padding and sibling-skill references that rely on fragile out-of-bundle relative paths.
Suggestions
Trim the motivational intro line ("This is a lane gstack has zero of — and one where a single missed clause costs more than a year of engineering") — it adds no guidance.
Make the sibling-skill references robust: the ../../../c-level-advisor/... relative paths do not resolve from this skill's directory; use stable skill/agent identifiers or move the scanner scripts into this skill's bundle.
Add an explicit validation step to the workflow (e.g., "confirm each top-3 issue appears in the Issues table and drives the verdict") to close the feedback loop.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Mostly lean: terse bullets with specific thresholds ("Standard cap: 12 months of fees", "Notice period (30 / 60 / 90 days)", "4-year, 1-year cliff") and no re-explanation of concepts Claude already knows. Not 5 because the intro contains motivational padding ("This is a lane gstack has zero of — and one where a single missed clause costs more than a year of engineering") that could be trimmed. | 4 / 5 |
Actionability | Fully executable for an instruction-style skill: a copy-paste-ready output template with worked example rows ("uncapped IP indemnity | HIGH | Cap at fees paid, mutual"), two concrete python commands with file arguments, and specific numeric checkpoints for every question. Covers the common cases without gaps. | 5 / 5 |
Workflow Clarity | Clear four-step sequence (read end to end → run the six questions → identify top-3 issues → apply the verdict) with the six questions serving as an explicit checklist and step 3 acting as a checkpoint. Not 5 because there is no explicit validation or feedback step (e.g., confirming the top-3 issues are reflected in the verdict); not 3 because the sequence and checkpoints are present and this is a review, not a destructive or batch operation, so the cap does not apply. | 4 / 5 |
Progressive Disclosure | Well-sectioned overview that appropriately pushes detail (contract scanner, term-sheet analyzer, three references) to a linked sibling skill, one level deep and clearly signaled under "Workflow Integration" and "Related". Not 5 because the relative paths (../../../c-level-advisor/...) do not resolve within this skill's bundle, making navigation fragile; not 3 because the structure and signaling are otherwise good. | 4 / 5 |
Total | 17 / 20 Passed |