Content
68%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 lean, well-structured checklist skill with concrete benchmarks, executable tool commands, and a complete output template. Its weakest point is the workflow: steps are listed but lack any verification checkpoint, and the external tool integration omits the term sheet analyzer's expected input format.
Suggestions
Add an explicit checkpoint to the Workflow (e.g., a step 3.5: 'Re-scan the contract for each of the six questions and confirm none was skipped before identifying the top-3 issues') so the sequence includes a validation loop, raising workflow clarity.
Specify the input format for term_sheet_analyzer.py (a short example of the expected term_sheet.json structure) so the command is fully executable without inspecting the external skill.
Trim non-earning tokens: the 'gstack' motivational sentence, the 'Since v2.5.1' lineage note, and the version footer add context-window cost without improving execution.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is efficient — terse checklists ("Standard cap: 12 months of fees", "409A current?") that assume Claude knows what HIPAA or GDPR are — but a few tokens don't earn their place ("This is a lane gstack has zero of…", "Since v2.5.1", the "Version: 1.0.0" footer), matching anchor 4 rather than 5. | 4 / 5 |
Actionability | Guidance is mostly executable: concrete numeric benchmarks ("12 months of fees", "30 / 60 / 90 days", "4-year, 1-year cliff"), copy-paste Python commands, and a complete fill-in output template. Minor gaps keep it below anchor 5 — the term_sheet_analyzer.py input JSON format is never specified and the script paths live outside this bundle. | 4 / 5 |
Workflow Clarity | The workflow lists a clear 4-step sequence ("Read the contract end to end → Run the six questions → Identify the top-3 issues → Apply the verdict") but includes no checkpoints (e.g., confirming all six questions were addressed or that identified issues are grounded in contract text), matching anchor 3 rather than 4. This is a read-only review skill, so the destructive/batch validation cap does not apply. | 3 / 5 |
Progressive Disclosure | Good structure with well-labeled sections (When to Run, Six Questions, Workflow, Output Format, Routing, Related) and clearly signaled one-level-deep references to the backing skill's scripts and references; no bundle files exist in this skill's directory, so the ~130-line body appropriately carries the content itself, with only minor organization gaps (Related/routing material could be tightened) — anchor 4. | 4 / 5 |
Total | 15 / 20 Passed |