Content
67%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 well-organized, actionable instruction-only skill: ready-to-use message templates, concrete priority orderings per query type, and clear rate-limit handling. Its main costs are token duplication of the source list and a small amount of illustrative padding, plus a reliance on an external reference that is not part of the bundle.
Suggestions
Deduplicate the six-source capability list: keep the table under "Checking Available Sources" and have the connect-guidance template reference it ("You currently have [N] source(s) connected" can list them inline once) instead of repeating the full list.
Trim or repurpose the sample "Source Status" ASCII block in Source Health into a one-line instruction ("report per-source availability as a list"), since the block is illustrative rather than executable.
Make the vague prevention directive concrete, e.g. "skip a source when its tool prefix is absent or the query type's priority list places it last and earlier sources already answered".
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The six-source capability list is duplicated verbatim (the table under "Checking Available Sources" and again inside the connect-guidance template), and the sample "Source Status" ASCII block is illustrative padding — the body is mostly efficient but could be tightened. It avoids re-explaining concepts Claude already knows, keeping it above a 2. | 3 / 5 |
Actionability | Copy-paste-ready user message templates ("[Tool name] isn't currently connected. To add it: 1. Open your MCP settings..."), concrete per-query-type priority orderings, and numbered connection steps give mostly executable guidance; a few directives like "check if the source is likely to have relevant results before querying" stay vague, preventing a 5. | 4 / 5 |
Workflow Clarity | Sequences are clearly ordered (connect: 3 steps; rate limiting: detection → handling → prevention with recovery guidance like "Do not retry immediately"), but there are no explicit validation checkpoints (e.g. confirming a newly added source is detected). No destructive or batch operations apply, so the 3-cap is not triggered. | 4 / 5 |
Progressive Disclosure | Good section structure in a single ~170-line file of appropriate size, with clearly signaled navigation. Not a 5 because the only reference points outside the skill bundle ("../../CONNECTORS.md", unverifiable from the skill directory) and there is no bundle to split the priority-table detail into. | 4 / 5 |
Total | 15 / 20 Passed |