Content
92%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.
An exemplary instruction-only skill body: fully actionable commands, explicit validation feedback loops, and clean progressive disclosure into a well-organized reference bundle. The only critique is occasional over-compression that trades clarity for brevity in the mention/delivery rules.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Extremely lean and dense with no padding or explanations of known concepts ('Missing commands permit a draft, never a DWS validation claim'), but a few passages (the @-mention and send-flag paragraphs) compress many rules into single sentences and there is minor redundancy between 'Preserve supplied facts and requested interactions' and 'Do not silently remove requested data, components, or interactions'. | 4 / 5 |
Actionability | Fully executable throughout: concrete dws aicard explain/lint/template/preview commands with flags, a copy-paste JSON message array example, and specific chat send/update commands with exit-code and flag details. | 5 / 5 |
Workflow Clarity | The intent-routing and scenario tables give clear sequences, with explicit validation checkpoints ('Check exit status, data.valid, and data.preflight.valid'), a fix-and-rerun feedback loop ('Fix errors at their diagnostic pointers and rerun the selected check'), and dry-run guidance for risky sends. | 5 / 5 |
Progressive Disclosure | The body is a concise overview routing by intent to one-level-deep, well-signaled references (design, patterns, index, protocol, templates); every cited path and anchor (e.g. #image-reconstruction, #root-surface, #short-notifications, #examples-and-replay) exists in the bundle, with explicit 'read only the relevant reference' guidance. | 5 / 5 |
Total | 19 / 20 Passed |