Content
22%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 content-free shell: it repeats the frontmatter description three times with slight rewording, provides no instructions, workflows, examples, or bundled reference files, and defers everything to an external repository link. Only its brevity and section headings keep it from scoring lower on conciseness and structure.
Suggestions
Replace the circular boilerplate in Instructions with concrete guidance: what to ask when a user reports symptoms, what format to log symptoms in, and how to analyze the medical information the skill claims to handle.
Define an explicit step-by-step workflow (e.g., gather symptom details → record in a structured log → analyze patterns → suggest next steps) with validation checkpoints such as directing serious symptoms to a healthcare professional.
Bundle the real content locally (e.g., references/symptom-tracking.md, references/wellness-checklists.md) instead of deferring entirely to an external GitHub repo that will not be available at runtime.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is short in absolute terms, but the same sentence is repeated verbatim across Overview, When to Use, and Instructions, and the Instructions section adds zero information ('This skill provides guidance and patterns for...'). This is unnecessary duplication that could be collapsed to a single statement — anchor 3 ('could be tightened'). Not 2 because there are no padded concept explanations; not 4-5 because the triplicate repetition means not every token earns its place. | 3 / 5 |
Actionability | The Instructions section is entirely abstract — 'This skill provides guidance and patterns for a health assistant skill' only describes rather than instructs, with no code, commands, steps, or concrete examples anywhere in the body. This matches anchor 1 exactly; it does not reach anchor 2 because there is not even a high-level hint of specific steps to execute. | 1 / 5 |
Workflow Clarity | No steps or sequence of any kind appear in the body; it does not define what to do when a user reports a symptom or how symptom tracking works. Anchor 1 ('steps missing or incoherent; no sequence') fits. The simple-skill exception cannot apply because there is no clear single action either — the body is operationally empty. | 1 / 5 |
Progressive Disclosure | Sections (Overview, When to Use, Instructions) are organized and the source-repo link is one level deep, but all substantive content is deferred to an external non-bundled GitHub URL rather than bundled reference files — the body is a shell. Anchor 3 fits ('some structure... could be better organized'). The under-50-lines exception's premise fails because the skill clearly relies on external material it does not bundle. | 3 / 5 |
Total | 8 / 20 Passed |