Content
85%Reviews 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, actionable, and uses clean one-level progressive disclosure with verified reference files. Its only weakness is mild verbosity from restating the gross/net headline-return framing in multiple sections.
Suggestions
State the gross/net headline-return labeling rule once (e.g., in Review Priorities) and reference it from the Workflow and Overview instead of restating it in each section.
Tighten the Review Priorities prose into tighter bullet fragments to reduce token overhead while keeping the field lists.
Consider folding the Fallback Behavior's 'ask the LP for the reporting package' guidance into the Workflow's Chronograph-unavailable branch to avoid near-duplicate instructions.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense and assumes domain competence, but the gross/net framing and headline-return guidance are restated across Overview, step 3, and Review Priorities, so it could be tightened to earn a 3. | 2 / 3 |
Actionability | As an instruction-only skill it gives highly specific guidance — named metrics (TVPI/DPI/RVPI, net IRR), explicit labeling rules, and a concrete output shape with 'three to eight prioritized questions' — rather than vague direction. | 3 / 3 |
Workflow Clarity | The eight-step workflow is clearly sequenced with explicit checkpoints: cite sources, separate facts from interpretation, label inferred items as analyst prompts, and list missing inputs on partial data. | 3 / 3 |
Progressive Disclosure | SKILL.md stays an overview and signals three real one-level-deep references (checklist, template, question-bank), with detailed material appropriately split into those files and no nested reference chains. | 3 / 3 |
Total | 11 / 12 Passed |