Content
82%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 well-organized, token-efficient overview body with concrete invocation examples and clear report-writing steps. Weaknesses are minor: an output field (`citations`) is referenced but never documented, no verification checkpoints before writing the report, and the primary detail reference lives outside the skill bundle at an unresolvable relative path.
Suggestions
Document the `citations` field in the output-structure sketch so step 6 ('Cite with [N] format from `citations`') has a visible counterpart in the brief example.
Move or copy the synthesis rules into the skill's own bundle (e.g., references/synthesis.md) and update the link so the reference resolves within the skill directory.
Add one explicit checkpoint before report writing, e.g., 'If sections.gaps covers a core sub-query, note the limitation up front and, if depth allows, rerun with a rephrased sub-query.'
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and dense with earned tokens: a compact JSON quick-reference, a depth-level table with concrete counts/times, a minimal output-structure sketch, six terse report-writing steps, and anti-pattern bullets. It explains nothing Claude already knows; the only soft spot is the one-line promotional tagline ('Beats chaining `search` + `fetch` manually'), which still conveys decision-relevant information. Matches anchor 5. | 5 / 5 |
Actionability | Five copy-paste JSON invocations cover the common cases (standard, comprehensive, quick, domain-scoped, custom schema) and the depth table gives exact parameters — mostly executable guidance. Minor gap: step 6 says 'Cite with [N] format from `citations`' but the documented output structure never shows a `citations` field, leaving the report-writer without the referenced detail. Anchor 4, not 5, due to this mismatch. | 4 / 5 |
Workflow Clarity | The report-writing workflow is a clear numbered sequence (start with key_findings, use cross_references for reliable claims, sections from topics, comparison table, note gaps, cite). The skill is non-destructive read-only research, so the missing-validation cap does not apply, but there are no explicit checkpoints (e.g., how to verify coverage before writing or handle a sparse brief beyond listing gaps). Anchor 4: clear sequence with most checkpoints, minor gaps. | 4 / 5 |
Progressive Disclosure | Good structure: the body stays a concise overview and delegates synthesis detail to a well-signaled one-level reference ('[wigolo/rules/synthesis.md](../wigolo/rules/synthesis.md)'), with See Also pointers to sibling skills. However, the reference path points outside the skill's own directory (../wigolo/rules/) and no such file exists in this skill's bundle (no references/, scripts/, or assets/ directories present), so navigation relies on an unverifiable external path. Anchor 4: good structure, references mostly clear, minor organization gaps. | 4 / 5 |
Total | 17 / 20 Passed |