Content
46%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This skill is comprehensive in coverage but suffers from significant verbosity and redundancy. It provides good executable code examples across many SymPy domains, but repeats patterns (especially lambdify), explains concepts Claude already knows, and inlines content that should live in the referenced files. The missing bundle files undermine the progressive disclosure structure that was designed.
Suggestions
Cut the content by ~50%: remove the 'When to Use This Skill' list, 'Getting Started Examples' (duplicates earlier sections), 'Additional Resources' (external URLs), and 'Integration with Scientific Workflows' (boilerplate Claude knows). Keep one canonical example per capability.
Remove duplicate lambdify examples — it appears in sections 7, Best Practices #5, Pattern 2, Integration with NumPy, and Getting Started Example 5. Show it once in the code generation section.
Either provide the referenced bundle files (core-capabilities.md, matrices-linear-algebra.md, etc.) or remove the references and keep the skill self-contained. Currently the references point to nonexistent files.
Fix incomplete imports in physics examples: add `from sympy import cos` and `from sympy.physics.quantum import Operator` to make the code actually executable.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is significantly verbose at ~350+ lines. It over-explains concepts Claude already knows (e.g., 'SymPy is a Python library for symbolic mathematics'), includes a lengthy 'When to Use This Skill' section listing obvious use cases, repeats similar code patterns multiple times (lambdify appears 4+ times), and the 'Getting Started Examples' section largely duplicates code already shown in earlier sections. The 'Integration with Scientific Workflows' section is mostly boilerplate that Claude would know. | 2 / 5 |
Actionability | The skill provides numerous concrete, executable code examples covering symbols, calculus, solving, matrices, code generation, and more. Most examples are copy-paste ready with correct imports. Minor gaps include the physics/mechanics examples which have incomplete imports (missing `cos` import, missing `Operator` import in quantum example) and some patterns that are more illustrative than fully executable. | 4 / 5 |
Workflow Clarity | The 'Solve and Verify' pattern includes a validation step (asserting solutions equal zero), and the 'Symbolic to Numeric Pipeline' shows a clear sequence. However, most of the skill presents isolated code snippets rather than workflows. The troubleshooting section helps but there are no explicit feedback loops for error recovery in the main workflow patterns. For a non-destructive skill this is adequate but not strong. | 3 / 5 |
Progressive Disclosure | The skill references five separate reference files (core-capabilities.md, matrices-linear-algebra.md, etc.) with clear loading guidance, which is good structure in principle. However, no bundle files are provided, meaning all those references are broken. Additionally, the main SKILL.md inlines extensive content that duplicates what the reference files should contain, creating a bloated overview that defeats the purpose of progressive disclosure. The 'Reference Files Structure' section is well-organized but the actual content organization doesn't follow through. | 2 / 5 |
Total | 11 / 20 Passed |