Content
85%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
A well-structured, actionable workflow skill with clear sequencing, explicit error handling and validation checks, and clean one-level-deep progressive disclosure. The main weakness is redundancy: several cautions and the 'do not hard-code tool names' directive are restated across multiple sections.
Suggestions
Consolidate the repeated cautions: state 'Do not hard-code tool names; read live MCP tool descriptions' once in the Chronograph MCP Usage section and remove the duplications in Overview and the custom-metrics bullet.
Merge the net-vs-gross guidance (currently in Chronograph MCP Usage, Error Handling, and Guardrails) into a single authoritative statement to remove verbatim repetition.
Centralize the 'future commitments are user-provided planning assumptions, not Chronograph actuals' rule — it currently appears in Workflow, Planning Mode, Error Handling, and Guardrails.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Mostly efficient and assumes domain competence (no definitions of NAV/IRR/MOIC/DPI/RVPI or the Takahashi-Alexander model), but several directives recur verbatim across sections — 'Do not hard-code tool names' (lines 14, 38, 44), the net-vs-gross caution (lines 40, 54), and the future-commitments-not-from-Chronograph note (lines 18, 53, 85) — which is the 'could be tightened' anchor rather than the lean top anchor. | 2 / 3 |
Actionability | Gives concrete, executable guidance despite being instruction-only: a 9-step sequence naming exact metrics to pull, named reference files for methodology and MCP mapping, a specific Excel workbook shape (Inputs/Fund Forecast/Portfolio Summary/Checks), and explicit reconciliation checks; absence of code is justified for this dynamic-MCP skill, per the instruction-only scoring note. | 3 / 3 |
Workflow Clarity | Clear numbered 9-step sequence plus a separate Planning Mode breakdown and an Error Handling section with explicit recovery actions (ask the user on entity-not-found / currency ambiguity; stop vs. proceed on MCP-not-connected by mode), and the Output Standards provide explicit reconciliation validation; not 2 because checkpoints are present, not merely implied. | 3 / 3 |
Progressive Disclosure | SKILL.md is an overview that signals two real, one-level-deep references — `references/model-methodology.md` (methodology/formulas) and `references/chronograph-mcp-map.md` (tool mapping) — both of which exist in the bundle, with a closing Reference Files index; content is appropriately split rather than inlined. | 3 / 3 |
Total | 11 / 12 Passed |