Content
67%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 dense, highly actionable reference: complete executable implementations for all eight capabilities, three end-to-end workflows, and useful troubleshooting/best-practices sections with a clear gating sequence. Its major structural flaw is that it never mentions the four bundled CLI scripts that duplicate and extend the inlined code, leaving the skill's most detailed materials undiscoverable and the body a monolith of inlined implementations.
Suggestions
Reference the bundled scripts from the body, e.g. a 'CLI Tools' section: gate_fcs.py for batch gating, immunophenotype.py for panel analysis, cfse_proliferation.py and cell_cycle.py for those assays, with their usage examples.
Move the full function implementations out of SKILL.md into the scripts (or a references/ file) and keep a concise Quick Start plus per-capability entry points, so the body becomes an overview with one-level-deep references.
Fix minor executability gaps: use `matplotlib.colors.LogNorm()`, pass channel indices/variables explicitly in the apoptosis example instead of hardcoded values, and make each workflow self-contained or clearly note its prerequisites.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense with executable code and light on prose padding, assuming Claude's competence (no explanations of what flow cytometry or the libraries are). Minor over-explanation remains: the Overview paragraph restates the frontmatter description nearly verbatim, and the 'When to Use' bullet list partially duplicates it. Matches 'efficient; minor instances that could be trimmed'. | 4 / 5 |
Actionability | Concrete, executable functions cover every capability (compensation parsing/application, rect/polygon/GMM gating, immunophenotyping, CFSE, Dean-Jett-Fox fitting, Annexin V quadrants, plotting) plus three end-to-end workflows and an install command. Minor gaps keep it below fully copy-paste ready: snippets rely on variables defined in earlier sections (e.g. `singlet_gate`, `channel_names`), the apoptosis example hardcodes `annexin_idx=3`, and `plt.matplotlib.colors.LogNorm()` is a nonstandard path (should be `matplotlib.colors.LogNorm`). | 4 / 5 |
Workflow Clarity | Multi-step processes are clearly sequenced: Best Practices prescribes 'debris exclusion first, then singlets, then viability, then markers', the gating code prints event counts after each step (feedback), and Troubleshooting gives concrete recovery fixes for failure modes; the cell-cycle fit even has a try/except error return. Minor validation gaps: the three Workflow examples lack explicit verification checkpoints (e.g. checking event counts or peak quality before proceeding), though Best Practices covers quality checks. Analysis is read-only, so the destructive/batch cap does not apply. | 4 / 5 |
Progressive Disclosure | The bundle contains four substantial standalone CLI tools in scripts/ (gate_fcs.py, immunophenotype.py, cfse_proliferation.py, cell_cycle.py; ~68 KB, 28 functions) that duplicate and extend the inlined code, yet SKILL.md never references them anywhere — zero referenced paths against a populated bundle. The body inlines ~500 lines of implementations that clearly belong in (and already exist as) separate files, matching the level-2 anchor 'content that clearly belongs in separate files is inlined'. Not level 3 because references aren't merely unclear — they are entirely absent and the duplication is total; not level 1 because the body itself is well-sectioned and internally navigable. | 2 / 5 |
Total | 14 / 20 Passed |