Content
57%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 highly actionable and rich in correct, executable Protocol API v2 examples, but it behaves as a monolithic API tutorial: it duplicates the existing reference file's content, buries its bundle pointers at the end, and offers no concrete simulate/validate step before robot execution. Restructuring around the existing reference and adding an explicit simulation feedback loop would lift the three dimensions stuck at 3.
Suggestions
Add an explicit validation step with a concrete command, e.g. 'Verify before robot execution: opentrons_simulate protocol.py must pass with no errors; fix and re-run until clean', before the complete protocol patterns.
Move the per-method API documentation (sections 4-7: well access, location methods, module control, execution control) into references/api_reference.md, keeping SKILL.md as an overview with a quick-start example and clearly signaled pointers near the top.
Surface the scripts/ templates with named links (e.g. 'Serial dilution: scripts/serial_dilution_template.py') in a Resources section placed earlier in the body, rather than a single unlisted pointer at the end.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is mostly dense, executable code with little concept padding, but sections 3-7 (well access, location methods, module control, execution control) systematically duplicate the method tables already in references/api_reference.md, which is more than minor over-inclusion. | 3 / 5 |
Actionability | Real, executable Protocol API v2 code throughout, including three complete run() protocols (serial dilution, plate replication, PCR setup) covering common cases. Minor gaps keep it from 5: many fragments depend on undefined variables and the plate-reader snippet uses 'plate_reader' without ever loading it. | 4 / 5 |
Workflow Clarity | Protocols are well sequenced and 'Simulate first' appears as a best practice, but there is no concrete simulation command, no validation checkpoint, and no validate-fix-retry feedback loop for protocols that batch-pipette on physical hardware, so the batch-operations cap applies. | 3 / 5 |
Progressive Disclosure | A real bundle exists (references/api_reference.md and three scripts/ templates) and both are referenced, but the pointers are buried in a trailing 'Resources' section while API-reference-grade content that belongs in the reference file is inlined in SKILL.md instead. | 3 / 5 |
Total | 13 / 20 Passed |