Content
38%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 reads like an auto-generated template with the description pasted in twice, duplicated sections, and boilerplate governance text, wrapped around a genuinely useful but unverifiable API walkthrough. Its most serious defects are that every referenced bundle file (scripts/, references/) is missing from the package and the code examples use inconsistent method names against that nonexistent API.
Suggestions
Remove the duplicated description pastes (the 'Use this skill when the task needs...' and 'Key Features' bullets), merge the two 'When to Use' sections and the two 'References' headings, and drop the generic template boilerplate (Output Requirements / Response Template / Input Validation) that adds nothing specific to citation mapping — this alone would cut the file roughly in half.
Ship the referenced bundle files (`scripts/main.py`, `references/guide.md`, `references/examples/`, `references/api-docs/`, `references/audit-reference.md`) or delete the references; broken paths are worse than none. Move the four Core Capabilities API walkthroughs into `references/api-docs/` and keep only a Quick Start in SKILL.md.
Reconcile the API surface in the code examples — `identify_research_clusters` vs `detect_clusters` and `visualize_network` vs `visualize` are inconsistent — and fix the 'See `## Workflow` above' pointer in Implementation Details, which points to a section that appears below it. Add a concrete validation step (e.g., check node/edge counts or that the output JSON parses) after network construction.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is noticeably padded: the frontmatter description is pasted verbatim twice ("Use this skill when the task needs Use when identifying seminal papers..." and again under Key Features), there are two competing trigger sections ("## When to Use" and "## When to Use This Skill"), two "## References" headings, and generic template boilerplate ("Output Requirements", "Response Template", "Input Validation", "Quality Checklist") that adds no citation-mapping-specific value. This matches the anchor for several unnecessary padded sections; it is above a 1 only because it does not explain basic concepts Claude already knows. | 2 / 5 |
Actionability | There is concrete-looking guidance (bash commands like `python -m py_compile scripts/main.py` and `python scripts/main.py --seed-pmid 12345678 --depth 2 ... --visualize`, plus worked Python examples), but the Python examples target an API that does not exist in the bundle and are internally inconsistent — `mapper.identify_research_clusters(...)` in Quick Start vs `mapper.detect_clusters(...)` in Core Capabilities, and `visualize_network(...)` vs `visualize(...)` — making them effectively pseudocode. This is the anchor for 'some concrete guidance but incomplete / pseudocode instead of executable code', not a 2, since specific commands and parameter values are given throughout. | 3 / 5 |
Workflow Clarity | Steps exist (a 4-step Example run plan, a 5-step ## Workflow, and an Error Handling fallback), but validation checkpoints are generic ('validate the request matches the documented scope') rather than verifying actual outputs, and the structure is muddled: three competing step lists, a 'See `## Workflow` above' pointer in Implementation Details that actually points to a section below, and no checkpoint that validates the produced network. This fits 'steps listed but validation gaps; checkpoints missing or implicit', not a 4. | 3 / 5 |
Progressive Disclosure | The body references `scripts/main.py`, `references/guide.md`, `references/examples/`, `references/api-docs/`, and `references/audit-reference.md`, yet none of these files exist in the bundle, so every reference is broken. Additionally, four subsections of API-style usage examples (Core Capabilities) are inlined directly in SKILL.md — content that the skill itself says belongs in `references/api-docs/`. Broken references plus inlined reference material matches the 'minimal structure / references buried or nonexistent' anchor; it is not a 3 because the navigation that exists does not resolve to real files. | 2 / 5 |
Total | 10 / 20 Passed |