Content
71%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.
A well-structured, highly actionable reference skill whose real value lies in version-specific caveats, scientific-plotting checks, and verified one-level-deep bundle references. Its main weakness is token weight: several sections re-teach baseline matplotlib knowledge Claude already has, and the export-validation guidance, while present, is scattered rather than sequenced as an explicit workflow.
Suggestions
Trim or pointer-ify sections that restate knowledge Claude already has — "The Matplotlib Hierarchy", "Two Interfaces", the basic plot-type one-liners, and "Integration with Other Tools" — keeping the body focused on the 3.11-specific changes and caveats that genuinely add value.
Consolidate the scattered export guidance into one ordered workflow (create → save with fig.savefig → inspect the exported file → fix and re-export) so the validation feedback loop is explicit rather than distributed across sections.
Reduce "Additional Resources" and the basic styling/rcParams listings to a few lines or move them into references/styling_guide.md to reclaim token budget.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Much of the body is dense, non-obvious curation (3.11 API changes, layout-engine interactions, savefig caveats, scientific checks), but sections like "The Matplotlib Hierarchy", "Two Interfaces" with trivial examples, the basic plot-type one-liners, "Integration with Other Tools", and "Additional Resources" restate knowledge Claude already has. Mostly efficient, but a meaningful amount of tutorial-level explanation could be trimmed or pushed to references. | 3 / 5 |
Actionability | Concrete, executable guidance throughout: copy-paste install commands ("uv add \"matplotlib==3.11.2\" numpy scipy") and runnable script invocations ("MPLBACKEND=Agg uv run --isolated ... scripts/plot_template.py --no-show --output plot.png"), plus specific API calls with parameter explanations. Not a 5 because most code fragments use undefined named data (x, y, X, Y, Z, data); this is honestly disclosed ("Fragment examples assume imports and named data; adapt and validate them before use") but still leaves a minor gap versus copy-paste-ready. | 4 / 5 |
Workflow Clarity | The setup → create → save → inspect sequence is present with explicit checkpoints: "Inspect the actual exported file at its final size for clipped labels, missing glyphs, contrast, and readable legends", "Save with fig.savefig before show, then plt.close(fig) in batch loops", and "adapt and validate them before use". Not a 5 because this guidance is distributed across topic sections rather than presented as one explicit ordered procedure with feedback loops, so the sequence must be assembled by the reader. | 4 / 5 |
Progressive Disclosure | The body is an overview with clearly signaled, one-level-deep references that all exist in the bundle: inline pointers ("For comprehensive plot type examples and variations, refer to `references/plot_types.md`"; "For detailed styling options and colormap guidelines, see `references/styling_guide.md`"), a consolidated "Detailed References" section listing all four reference files, and documented usage commands for both bundled scripts. Content is appropriately split and easy to navigate. | 5 / 5 |
Total | 16 / 20 Passed |