Content
27%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This skill is excessively verbose and repetitive, with circular internal references, duplicated content across sections, and large amounts of generic boilerplate (security checklists, lifecycle status, evaluation criteria) that don't help Claude perform the specific task of checking figure references. The core actionable content—run `python scripts/main.py --manuscript paper.docx` to check figure references—is buried under layers of process documentation. The skill would benefit enormously from being reduced to roughly 20-30 lines focusing on the actual command, parameters, and expected output.
Suggestions
Remove all circular self-references ('See ## Features above') and consolidate duplicated content into single authoritative sections—the py_compile command appears 3 times, the script path appears 5+ times.
Delete generic boilerplate sections (Security Checklist, Lifecycle Status, Evaluation Criteria, Risk Assessment, Response Template) that don't provide figure-reference-checker-specific guidance Claude needs.
Add a concrete example showing actual input and expected output—e.g., a sample manuscript snippet with a mismatched figure reference and what the tool's output looks like.
Collapse the entire skill to: purpose (1 line), usage command with parameters, example input/output, error handling for the specific tool, and a reference link—targeting under 50 lines total.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Extremely verbose and repetitive. Multiple sections reference each other circularly ('See ## Features above', 'See ## Prerequisites above', 'See ## Usage above'). The same information (e.g., py_compile command, script path, workflow steps) is repeated across multiple sections. Contains boilerplate sections (Lifecycle Status, Security Checklist, Evaluation Criteria) that add no actionable value for Claude. The skill explains obvious concepts and includes generic templates that Claude already knows. | 1 / 3 |
Actionability | There is some concrete guidance—the CLI command `python scripts/main.py --manuscript paper.docx` is specific and the parameter table is clear. However, much of the content is generic process description rather than executable instruction. The workflow steps are abstract ('Confirm the user objective', 'Validate that the request matches the documented scope') rather than concrete commands or code. The actual features (orphaned references, label consistency, missing citations) lack implementation details. | 2 / 3 |
Workflow Clarity | There is a numbered workflow with steps and error handling guidance, including a fallback path for failures. However, the workflow is generic and process-oriented rather than specific to figure reference checking. There are no validation checkpoints tied to the actual output (e.g., 'verify the report contains all referenced figures'). The error handling section mentions fallbacks but doesn't provide concrete recovery commands. | 2 / 3 |
Progressive Disclosure | The document contains circular self-references ('See ## Features above for related details' within the Key Features section itself), which is confusing and adds no navigational value. References to `references/audit-reference.md` and `scripts/main.py` exist but no bundle files are provided to verify them. The content is a monolithic wall of repetitive sections that could be dramatically condensed, with much boilerplate that should either be removed or split into separate files. | 1 / 3 |
Total | 6 / 12 Passed |