Content
66%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's core is strong: a well-sequenced workflow with real validation and feedback loops, and configuration guidance that matches the bundled script exactly. Its weakness is token efficiency — four overlapping rule sections repeat the same structural advice, and roughly half the tail of the file is generic boilerplate that dilutes the actionable content.
Suggestions
Collapse 'Content Richness Requirements', 'Information Extraction Priority', 'Paper Input Enhancement Rules', and 'Prohibited Behaviors' into one deduplicated 'Structuring rules' section (or move them into references/input-patterns.md alongside the templates), since they repeatedly state the same keep-details/don't-compress guidance.
Delete or drastically trim the generic boilerplate sections ('When to Use', 'When Not to Use', 'Required Inputs', 'Output Contract', 'Failure Handling', 'User Checkpoints', 'Input Validation', 'Quick Validation') — statements like 'Use this skill when the user explicitly needs to perform the core task of mindmap-html-generator' are circular and consume context without adding skill-specific guidance.
Add a short filled-in example of the CONFIG block (input_path, output_html_path, page_title set to real values) in step 3 so the configure-then-run path is fully copy-paste ready, closing the actionability gap.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Four overlapping rule sections ("Content Richness Requirements", "Information Extraction Priority", "Paper Input Enhancement Rules", "Prohibited Behaviors") repeat the same guidance multiple ways ("conciseness does not mean deleting details", "do not sacrifice information density for brevity", "do not omit key numbers, key steps..."), and roughly 60 lines of generic boilerplate ("When to Use" — which circularly says to "perform the core task of mindmap-html-generator" — plus "Output Contract", "User Checkpoints", "Input Validation", "Quick Validation") add no skill-specific value. This goes beyond the 'some unnecessary explanation' of anchor 3 and matches anchor 2 (noticeably verbose, several padded sections), though it avoids anchor 1 since it never explains background concepts Claude already knows. | 2 / 5 |
Actionability | The body gives the exact run command (`python scripts/generate_mindmap_html.py`), CONFIG option names that match the bundled script (input_path, output_html_path, auto_bold_levels, plain_text_root_max_children, and MARKMAP_INIT_OPTIONS["colorFreezeLevel"]), and concrete structural rules (depth 2-4 levels, at least 6 level-1 branches, 3-6 level-2 nodes). The main gap is that no filled-in CONFIG example is shown, so it sits at anchor 4 (mostly executable, minor gaps) rather than the fully copy-paste-ready anchor 5. | 4 / 5 |
Workflow Clarity | The workflow is a clearly sequenced 5-step process (determine input source, organize structure, configure script, run script, validate results) with explicit validation checks ("Confirm only level-1 and level-2 headings remain bold") and error-recovery feedback loops ("If nodes are too long, go back to compress input text, then re-run the script"; install Node.js if npx is missing), matching anchor 5. It does not fall to anchor 4 because checkpoints and retry loops are fully present rather than having minor gaps. | 5 / 5 |
Progressive Disclosure | Both referenced bundle files exist and are clearly signaled one level deep ("Read when deciding how to organize input — contains recommended structure templates and examples"), and the Resources section accurately describes the script and reference file. However, ~80 lines of generic policy boilerplate (When to Use / When Not to Use / Required Inputs / Output Contract / Failure Handling / User Checkpoints / Input Validation / Quick Validation) sit inline in SKILL.md where they should be trimmed or split out, so content placement has minor organization gaps — anchor 4 rather than the appropriately-split anchor 5. | 4 / 5 |
Total | 15 / 20 Passed |