Content
77%Weight 40%Scale 1-3Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
A highly actionable, well-sequenced planning skill with concrete model IDs, metrics, templates, and an explicit feasibility gate. Its weaknesses are minor verbosity from cross-domain analogies and a monolithic structure that does not use progressive disclosure across files.
Suggestions
Trim context-free analogies such as the collider/MFA comparison line to improve token efficiency.
Move the four Study Archetypes and/or the study_card.md template into one-level-deep reference files (e.g. references/archetypes.md) referenced from SKILL.md to apply progressive disclosure.
Consider tightening the Planning Inputs and default-organism tables by collapsing low-value 'Why' rationales into the main column.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Mostly efficient structured tables, metrics, and templates, but includes unnecessary asides such as 'This is the MFA analogue of choosing a collider process and parameter scan before generating events' that do not earn their tokens. | 2 / 3 |
Actionability | Provides concrete model IDs (iJO1366, iMM904, Recon3D, iNJ661), named products, a full study_card.md template, and specific required metrics; as an instruction-only planning skill, code absence is not penalized. | 3 / 3 |
Workflow Clarity | Clear sequence (inputs -> archetype -> feasibility gate -> study card -> per-stage guidance) with an explicit validation checkpoint ('Proceed only if total score is at least 18/25. Otherwise choose a simpler organism...') and a study-card checklist. | 3 / 3 |
Progressive Disclosure | Sections are well organized, but the body is an entirely monolithic ~225-line file with no bundle files; the four archetypes and the study-card template are inline content that could plausibly be split into one-level-deep references. | 2 / 3 |
Total | 10 / 12 Passed |