Content
63%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 lean, well-sequenced guide with a useful worked example, but it is undermined by bundle integrity problems: the referenced parser script is absent from the bundle and the existing reference file is never linked. Fixing the script reference and wiring up the reference file would lift both actionability and progressive disclosure.
Suggestions
Ship the referenced `parse_query.py` in a scripts/ directory (or replace the script step with inline parsing guidance), and show the exact invocation command.
Link references/query_components.md from the body (e.g., under Core Task or Procedure) so the attribute/price-pattern tables are discoverable, and drop the duplicated search-strategy tips from Notes.
Add an explicit verification step, e.g., "Before clicking, confirm each result's price against price_max/price_min" to sharpen the workflow's checkpoint.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The ~30-line body is efficient and assumes Claude's competence, with a concrete worked example. Not 5 because the "When to Use" section restates the frontmatter description and the Notes duplicate search-strategy tips that also live in references/query_components.md; not 3 since padding is minor. | 4 / 5 |
Actionability | The core step says "Execute the bundled script `parse_query.py`", but no scripts/ directory exists in the bundle and no invocation command is given, so the primary instruction cannot be executed as written. The worked example ("{'product': 'popcorn', 'attributes': ['gluten free'], 'price_max': 140.0}" → "search[gluten free popcorn]") is concrete, which keeps it at anchor 3 rather than 2. | 3 / 5 |
Workflow Clarity | A clear 3-step sequence (parse → review & refine → formulate search strategy) with a refine checkpoint and an explicit error-recovery fallback ("If the initial search yields no results, consider broadening the search..."). Not 5 because the checkpoint is vague ("use your judgment") and there is no explicit constraint-verification step. | 4 / 5 |
Progressive Disclosure | Sections are well organized, but scored against the actual bundle: references/query_components.md exists yet is never linked from the body, its search-strategy guidance is partially duplicated inline in Notes, and the referenced path `parse_query.py` does not resolve to any bundled file. Fits anchor 3 (references present but not clearly signaled) better than anchor 2 since the body itself is structured, not inlined bulk content. | 3 / 5 |
Total | 14 / 20 Passed |