Content
75%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 dense, well-organized instruction skill: concrete query templates, defined output formats, and sequenced workflows with verification habits throughout. Its main weaknesses are mildly over-dense prose in the discovery sections and underuse of the reference-file structure for detailed module content.
Suggestions
Tighten the 'Discover competitors' and 'Discover products through their distribution channels' prose into the same terse table/list style used by the SEO modules to reduce token cost without losing guidance.
Move the detailed research-type modules (keyword, content, backlink, pricing tables) into a reference file, keeping SKILL.md as a shorter overview that selects the module by scope.
Add a one-line worked example of an evidence register row (comparison unit, task, relationship, coverage, surface, URL, date, confidence) to make the abstract register spec directly executable.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is efficient overall — terse tables, imperative instructions, and no explanation of concepts Claude already knows — but the dense discovery and distribution-channel prose paragraphs could be tightened and the opening paragraph partially repeats the description, fitting the 4 anchor (minor trimming possible) rather than the 5 anchor where every token earns its place. | 4 / 5 |
Actionability | Guidance is mostly executable for a research skill: copy-paste query templates ("[brand] alternatives", "best [tool] for [audience]", site:competitor.com), named tools, a defined evidence-register field list, and a concrete report structure. It falls short of 5 because some positioning guidance remains abstract (e.g., "Use product comparisons to identify meaningful differences and claims the client can support") with no worked example. | 4 / 5 |
Workflow Clarity | Multi-step processes are clearly sequenced (numbered discovery steps, the article fetch workflow, Parse→Enrich→Build→Generate) with several verification checkpoints ("verify identity before adding candidates," "check the claimed capability at the source," "Leave uncertain relationships pending"), matching the 4 anchor; it is not 5 because there are no explicit error-recovery feedback loops, and no destructive/batch cap applies to read-only research. | 4 / 5 |
Progressive Disclosure | The body is well-sectioned and the one bundle reference (references/research-workflow.md) is real, clearly signaled, and one level deep ("Use [research workflow](references/research-workflow.md)…"), matching the 4 anchor for good structure with minor gaps; it is not 5 because substantial method detail (SEO module tables, distribution-channel prose) is inlined in SKILL.md rather than split into the existing reference structure. | 4 / 5 |
Total | 16 / 20 Passed |