Content
75%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, largely executable reference: real commands, verified tool behavior, and useful gotchas, with clear per-topic sequencing. The main weaknesses are unverifiable cross-skill fallback paths and small pockets of textbook chemistry that pad the token budget.
Suggestions
Fix or inline the fallbacks that reference the sibling skill (crystal_validator.py, chemistry_facts.py) — either bundle those scripts here, drop the references, or state clearly they require tooluniverse-organic-chemistry to be installed.
Trim the 'Verified live' paragraph to the workflow and gotcha (alias pairs, empty-result meaning) and drop the quoted example cell parameters, which add tokens without adding instruction.
Cut or compress the §5 equation list (ΔG, Arrhenius, Nernst, Clausius-Clapeyron) to just the 'COMPUTE DON'T ESTIMATE — write Python' directive, since Claude already knows these textbook formulas.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Mostly lean, high-signal content: ready-to-run commands, tool-call syntax, fallback paths, and verified gotchas ("an empty `data: []` with `count: 0` means no matching structures in COD"), matching anchor 4's "efficient; minor instances of over-explanation that could be trimmed". It is not anchor 5 because of textbook material Claude already knows (the §5 equation list ΔG/Arrhenius/Nernst/Clausius-Clapeyron, XeF2/XeF4/XeF6 geometries) and the padded "Verified live" paragraph quoting example cell parameters (a=9.7397, b=8.9174...), but these are minor and not the "several unnecessary explanations" of anchor 2-3 territory. | 4 / 5 |
Actionability | Mostly executable, copy-paste-ready guidance: full CLI invocations with all flags ("--type ksp_kf --ksp 5.3e-27 --kf 1.1e33 --stoich 1:3"), concrete JSON tool calls, and exact solver paths verified against the bundled scripts/equilibrium_solver.py — matching anchor 4. It falls short of anchor 5 because two fallback paths ("skills/tooluniverse-organic-chemistry/scripts/crystal_validator.py" and ".../chemistry_facts.py") point outside this bundle and those files are not present, so those commands are not verifiable as runnable here. | 4 / 5 |
Workflow Clarity | Each topic section is clearly sequenced: §1 gives a compute-then-verify workflow with a preferred tool, a fallback, a batch mode, and an explicit error-recovery checkpoint ("try relaxing filters... before concluding the compound isn't in COD"), and §6 pairs each command with its key parameters — matching anchor 4's "clear sequence with most checkpoints present; minor validation gaps". It is not anchor 5 because feedback loops are only implicit for most sections (no validate-then-fix-then-retry cycle beyond the COD gotcha), and no destructive/batch operation is involved, so the ≤3 cap does not apply. | 4 / 5 |
Progressive Disclosure | Good structure against the actual bundle: a single 116-line SKILL.md with clean section headers, a tool summary table, and the one bundled script (scripts/equilibrium_solver.py) referenced directly with runnable examples — matching anchor 4's "good structure; most content appropriately placed; minor organization gaps". It is not anchor 5 because the two cross-skill script references point at files not present in this bundle (a navigation/reliability gap), and the COD lookup detail (verified examples, alias pairs) is inline content that could live in a one-level-deep reference file. | 4 / 5 |
Total | 16 / 20 Passed |