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 usage material (commands, parameter table, preset table, output example) is strong and executable, and the workflow has sensible checkpoints including a fallback path. However, the body is heavily padded with duplicated commands, broken 'see above' template artifacts, and generic scaffolding sections, while its References section points to files that do not exist in the bundle.
Suggestions
Remove the template boilerplate — the broken 'See `## Features`/`## Usage`/`## Workflow` above' pointers, the duplicated py_compile and requirements.txt blocks, and the off-topic 'data analysis tasks / documented fallback path' bullets in When to Use.
Fix the progressive-disclosure problem: either create the `references/` files the References section promises (PyMOL techniques, ChimeraX lighting, colorblind palettes, journal guidelines) or delete the dangling References and requirements.txt mentions.
Cut the generic meta sections (Security Checklist, Risk Assessment, Lifecycle Status, Response Template, Output Requirements) or move them out of SKILL.md, keeping only what changes Claude's behavior for this task.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The ~320-line body is noticeably padded with template artifacts and duplication: "python -m py_compile scripts/main.py" appears three times, "pip install -r requirements.txt" appears in two separate prerequisite sections, and broken boilerplate lines like "See `## Features` above for related details." and "See `## Workflow` above for related details." add no value. Generic scaffolding sections (Security Checklist, Risk Assessment, Output Requirements, Response Template, Lifecycle Status) and off-topic 'When to Use' bullets about 'data analysis tasks that require explicit assumptions, bounded scope' pad the token budget well beyond what the task needs. This sits between score 2 ('several unnecessary explanations or padded sections') and score 3; the volume of duplicated and irrelevant boilerplate places it noticeably below the midpoint, and it is not score 1 because it never explains concepts Claude already knows — the parameter and preset tables are dense and useful. | 2 / 5 |
Actionability | The body provides fully executable, copy-paste-ready commands ("python scripts/main.py --pdb 1mbn --preset cover", multi-flag advanced examples) plus a complete parameter table with types/defaults and a preset comparison table, matching 'Mostly executable guidance; concrete code or commands with minor gaps'. It falls short of score 5 because the References section points to a `references/` directory that does not exist in the bundle, `requirements.txt` is referenced but missing, and the Example Usage includes an environment-specific path ("cd \"20260318/scientific-skills/Data Analytics/3d-molecule-ray-tracer\"") that will not resolve for users. It is above score 3 because nothing is pseudocode — every command shown is runnable. | 4 / 5 |
Workflow Clarity | The Workflow section gives a clear 5-step sequence with explicit checkpoints (confirm objective/constraints, validate scope and 'stop early if the task would require unsupported assumptions', fallback path on failure), reinforced by a Quick Check ("python -m py_compile scripts/main.py") and an Error Handling section stating 'report the failure point, summarize what still can be completed safely, and provide a manual fallback'. This matches 'Clear sequence with most checkpoints present; minor validation gaps' — the gap being that no step verifies the generated .pml/.cxc script actually runs in the target software (Evaluation Criteria lists this only as unchecked boxes). It is below score 5 because validation of the final artifact is implied rather than an explicit feedback loop in the workflow steps. | 4 / 5 |
Progressive Disclosure | Structure exists (a genuinely used, real script path `scripts/main.py`, plus parameter/preset tables), but the body is a ~320-line monolith that inlines generic meta-content (security checklists, lifecycle status, response templates) that belongs elsewhere or nowhere, and the References section signals four reference documents ('PyMOL-specific rendering techniques', 'ChimeraX lighting documentation', 'Colorblind-friendly palettes', 'Journal submission guidelines') that do not exist in the bundle — the `references/` directory is absent. This matches 'Some structure but could be better organized; references present but not clearly signaled; content that should be separate is inline'; it is above score 2 because real structure and a working script reference exist, but below score 4 because the flagged references are dangling. | 3 / 5 |
Total | 13 / 20 Passed |