Content
56%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 core PICOS extraction workflow (Steps 1-2 plus the JSON output format) is well-designed and executable, but it is buried in ~90 lines of generic template filler that adds token cost without task-specific value, and the peripheral sections contradict each other on validation. The description is strong; the body needs trimming and consistency fixes.
Suggestions
Cut the generic boilerplate sections (When to Use, Key Features, Dependencies, Implementation Details, Output Contract, Validation and Safety Rules, Failure Handling, Deterministic Output Rules, Completion Checklist) and keep only the PICOS workflow, the output format, and the when-not-to-use guidance — this alone would roughly halve the body.
Resolve the validation contradiction: "Validation Shortcut" says to run `python scripts/validate_skill.py --help` while "Quick Validation" says no script validation is required. Keep one, and remove references to inapplicable concepts (input/output paths, CONFIG blocks, thresholds) that do not apply to a title-only task.
Fix the navigation reference "See `## Usage` above" — the Usage section appears below that line in the file — and remove the duplicated top-half (lines up to "Validation Shortcut") in favor of the actual skill body that starts at "# Meta PICOS Generator".
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Roughly half the body is generic template boilerplate ("When to Use", "Key Features", "Dependencies", "Implementation Details", "Output Contract", "Validation and Safety Rules", "Failure Handling", "Completion Checklist", "Deterministic Output Rules") that restates generic discipline rather than anything specific to PICOS extraction — several padded sections. Not 1 because the core section is not explanatory filler and does not teach concepts Claude already knows. | 2 / 5 |
Actionability | The core instructions are executable: a worked title-cleaning example, explicit per-field extraction rules, a default rule ("If the type is not mentioned in the title, default to 'Randomized Controlled Trial'"), and an exact JSON output format. However, surrounding guidance is misleading or inapplicable — "Parameters to clarify first: input path, output path, scope filters, thresholds" and "Edit the in-file CONFIG block" suggest a file-based pipeline the title-only workflow does not have, and the validation commands reference a helper script that does no PICOS validation. This lands between anchor 3 and 4; the inapplicable directions keep it at 3. | 3 / 5 |
Workflow Clarity | The actual task workflow is a clearly sequenced two-step process (clean title → extract PICOS → emit JSON) with an example and explicit constraints. Not 5 because validation checkpoints are self-contradictory: "Validation Shortcut" instructs running `python scripts/validate_skill.py --help` while "Quick Validation" states "No local script validation step is required for this skill"; not 3 because the core sequence itself is coherent and complete for this non-destructive task. | 4 / 5 |
Progressive Disclosure | The single bundle reference (`scripts/validate_skill.py`) is a real, clearly signaled file, but the 165-line monolithic body mixes the core instructions with boilerplate that belongs in neither place, and navigation is broken: "See `## Usage` above" points to a section that actually appears later in the file. Some structure exists via headers, but organization could be materially better; not 2 because the sections are clearly labeled and the script path is not buried. | 3 / 5 |
Total | 12 / 20 Passed |