Content
23%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 severely bloated and largely non-actionable. It reads more like a product specification document than an executable skill for Claude, with extensive pseudocode that can't be run, explanations of obvious concepts, and massive inline content that should be split into supporting files. The referenced scripts and knowledge base files don't exist in the bundle, making most of the concrete guidance fictional.
Suggestions
Reduce the skill to under 100 lines by removing explanations of known concepts (OAuth, CSV, what Apple Health is), the FAQ section, extensibility notes, and performance optimization tips — none of these help Claude execute the skill.
Replace pseudocode JavaScript with actual executable code or concrete tool invocations (Read, Write, Grep, Glob) that Claude can perform, since those are the stated available tools.
Create actual bundle files for the referenced paths (scripts/, knowledge-base/index.md, mapping configs) or remove references to non-existent files.
Add real validation feedback loops: specify what to check after data import, what constitutes invalid data with concrete examples, and explicit recovery steps when validation fails.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Extremely verbose at ~500+ lines. Extensively explains concepts Claude already knows (what Apple Health is, how OAuth works, what CSV files are). Massive amounts of pseudocode that isn't executable, redundant output format examples, FAQ section explaining obvious things, and sections like 'Extensibility' and 'Performance Optimization' that add no actionable value. The skill could be reduced to ~20% of its size without losing meaningful information. | 1 / 5 |
Actionability | The code examples are pseudocode/illustrative JavaScript that isn't executable — functions like `identifySource()`, `readAppleHealthExport()`, `fetchFitbitData()` are referenced but never defined. The bash commands reference scripts (e.g., `python scripts/import_apple_health.py`) that don't exist in the bundle. The skill describes what should happen rather than providing concrete, executable steps Claude can actually perform. | 2 / 5 |
Workflow Clarity | Steps are listed in a logical sequence (identify intent → import data → validate → report), but validation steps are superficial pseudocode without real feedback loops. The validation function just returns a stub object. There's no explicit 'if validation fails, do X' recovery loop for the data import workflow — the validation step in 2.4 is illustrative only. For a skill involving data import (a potentially destructive/batch operation), the lack of real validation caps this at 3. | 3 / 5 |
Progressive Disclosure | The skill is a monolithic wall of content (~500+ lines) with no bundle files to support it. Content that should clearly be in separate files (integration guides, knowledge base index, output format templates, article mapping data) is all inlined. References to files like `knowledge-base/index.md`, `integrations/`, and `scripts/` directories don't exist in the bundle, making the references hollow. | 2 / 5 |
Total | 8 / 20 Passed |