Content
86%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 tight, operationally specific guide: it names the two tools, their key parameters and behavioral constraints, validation-before-rerun checkpoints for batch dispatch, and a concrete output contract. Its only notable gap is the absence of an example invocation and an explicit shortlist decision rule, which keeps actionability and workflow clarity just below full marks.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and assumes domain competence — no explanation of what MACE, DFT, or ML potentials are; every line carries operational facts like "Both tools reject `output_root` inside `input_dir`". It matches anchor 5: every token earns its place. | 5 / 5 |
Actionability | Concrete, tool-specific guidance throughout — named tools (`mace_relax_batch`, `mace_sp_batch`), parameters (`model`, `head`, `dispersion`, `relax_lattice`), and output artifacts (`batch_state_rel`, `batch_summary_rel`). It falls short of anchor 5 only in lacking an example invocation and an explicit shortlist decision rule. | 4 / 5 |
Workflow Clarity | Quick Start gives a clear 4-step sequence, and validation checkpoints are present for the batch operation ("inspect those before deciding to rerun", "Use collected evidence, not launch success alone"), so the batch cap at 3 does not apply. It stays at anchor 4 because the keep/drop shortlist criterion is demanded in the Output Contract but never actually specified. | 4 / 5 |
Progressive Disclosure | Under 50 lines with well-organized sections (Overview, Quick Start, Suggested tools, Workflow, Method-critical defaults, Output Contract, References) and no need for external reference files — the only cross-reference (`vasp-input-preparation`) points to another skill at one level of depth. Per the simple-skill guideline this matches anchor 5. | 5 / 5 |
Total | 18 / 20 Passed |