Content
67%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 technical core is excellent — executable commands for every input mode, a complete worked example, a thorough CLI reference, and a verification loop with concrete escalation steps. The score is dragged down by a large tail of generic, skill-agnostic boilerplate that wastes context and even contradicts the document's own validation guidance.
Suggestions
Delete or drastically condense the generic sections from 'When Not to Use' through 'Completion Checklist' (~85 lines of template boilerplate); keep only skill-specific guidance such as the existing verification workflow and failure handling for Puppeteer/browser errors.
Resolve the contradiction between the 'Validation Shortcut' section (run scripts/validate_skill.py --help) and the 'Quick Validation' section ('No local script validation step is required for this skill') — remove whichever does not reflect the real workflow.
Consider moving the full 40-line single-page HTML example into a references/ file (e.g., references/single-page-example.html) and keeping just the command and key CSS rules inline.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The core sections (When to Use, Example Usage, Implementation Details) are lean and earn their tokens, but roughly the last third of the body — 'When Not to Use', 'Required Inputs', 'Recommended Workflow', 'Output Contract', 'Validation and Safety Rules', 'Failure Handling', 'Quick Validation', 'Deterministic Output Rules', 'Completion Checklist' — is generic template boilerplate with lines like "Keep the output safe, reproducible, and within the documented scope at all times" that add no skill-specific value. This is several padded sections, matching the 'noticeably verbose' anchor rather than 'mostly efficient with some unnecessary explanation'. | 2 / 5 |
Actionability | Fully executable, copy-paste-ready guidance throughout: exact commands ('node assets/html-to-pdf.js input.html output.pdf', 'npm install'), a complete runnable HTML example for the single-page A4 pattern, a full CLI parameter table with defaults, and concrete recovery values ('--scale' e.g. '0.9', '0.8', '--margin=10mm'). Common cases (local file, URL, stdin, RTL, custom executable) are all covered by example. | 5 / 5 |
Workflow Clarity | The recommended verification workflow provides an explicit feedback loop: check 'No extra blank pages' / 'No clipped text', then iterate up to 5 attempts with specific fixes, escalating to HTML/CSS correction. However, the overall validation story has gaps: the top 'Validation Shortcut' (scripts/validate_skill.py --help) serves no clear purpose in the conversion flow, and the later 'Quick Validation' section states 'No local script validation step is required for this skill', directly contradicting it — so it sits below the fully coherent anchor 5. | 4 / 5 |
Progressive Disclosure | The bundle structure is used appropriately: the implementation script lives in assets/html-to-pdf.js and the validation helper in scripts/validate_skill.py, both referenced by exact, verified-existing paths, with no nested references. Sections are well-organized with clear headers, but ~85 lines of boilerplate (the generic governance sections, plus the 40-line HTML example that could live in a references file) keep it at 'minor organization gaps' rather than an ideal overview split. | 4 / 5 |
Total | 15 / 20 Passed |