Content
82%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.
The body is a high-quality, executable, well-sequenced analytical workflow with strong domain reasoning and concrete pitfalls baked into code. Tightening the longer prose sections and offloading the API reference to a bundled file would push it to the top tier.
Suggestions
Trim the multi-paragraph reasoning sections (Phase 3 'How to Reason' and Phase 6 synthesis) to bullet lists where possible to reduce token load.
Move the 'Programmatic Access (Beyond Tools)' REST code reference into a bundled references/ file and link to it from the body to improve progressive disclosure.
Add an explicit verification checkpoint (e.g., a final 'Validate the causal chain' checklist with the evidence tiers) before the synthesis output to strengthen workflow clarity.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Dense, competency-assuming content with tight code blocks and no concept padding, but several explanatory prose sections (target-gene reasoning, mechanistic synthesis, scenarios) could be trimmed slightly without losing value. | 4 / 5 |
Actionability | Fully executable code blocks with exact parameter names, documented return formats, concrete worked examples (rs7903146/TCF7L2/pancreas), and explicit parameter-pitfall callouts covering the common cases. | 5 / 5 |
Workflow Clarity | Six phases are clearly sequenced with fallbacks and evidence-graded confidence tiers, but explicit validate-then-act checkpoints are softer since this is analytical synthesis rather than a destructive/batch operation. | 4 / 5 |
Progressive Disclosure | No bundle files exist; the body is well-structured with clear section headers and a single one-level reference to a sibling skill, though the inline API-reference and programmatic-access content could be split into a separate file. | 4 / 5 |
Total | 17 / 20 Passed |