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.
A tight, well-organized single-file research skill with strong standards, security guidance for untrusted sources, and an explicit pre-delivery quality gate. The main gaps are the absence of a worked output example and a fix-and-retry loop behind the quality gate.
Suggestions
Add a short worked example of the output structure (e.g., a 5-line mock executive summary with a sourced number and a labeled estimate) so the default format is copy-paste concrete.
Extend the Quality Gate with one feedback line: what to do when a check fails (re-source the number, re-flag stale data, re-derive the recommendation) before delivering.
Name one or two concrete sourcing starting points per research mode (e.g., public filings, funding databases, analyst reports) to sharpen actionability for the collection steps.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and terse (~85 lines of compact bullets) with no padding or explanations of concepts Claude already knows; every line adds skill-specific guidance ('Produce research that supports decisions, not research theater'). It assumes Claude's competence throughout, matching the top anchor. | 5 / 5 |
Actionability | Concrete, specific guidance for an instruction-only skill: per-mode collection checklists ('fund size, stage, and typical check size', 'top-down estimates... bottom-up sanity checks'), a default output structure, and a pre-delivery quality gate. It falls short of fully executable guidance because it gives no worked example of the output format and no concrete sourcing method or search strategy. | 4 / 5 |
Workflow Clarity | The sequence is clear (activate -> research standards -> mode-specific collection -> output format -> quality gate) and the Quality Gate section is an explicit validation checklist before delivery. However, there is no feedback loop for what to do when the gate fails (fix and re-verify), which the top anchor expects. | 4 / 5 |
Progressive Disclosure | This is a single-file skill with no bundle files, and the content warrants none: each research mode is ~8 compact lines that belong inline. Sections are clearly headed (When to Activate, Research Standards, Untrusted Sources, Common Research Modes, Output Format, Quality Gate) making navigation trivial, matching the well-organized-structure standard. | 5 / 5 |
Total | 18 / 20 Passed |