Content
35%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 persona résumé: an expansive catalog of databases, techniques, and traits that assumes Claude knows nothing, while providing zero executable guidance — no example queries, no EXPLAIN workflows, no concrete tuning steps. It works only as a vague scope-setter; almost all of it could be cut or moved to references without loss.
Suggestions
Replace the Capabilities/Behavioral Traits/Knowledge Base catalogs with a small set of concrete workflows: an example query-optimization loop (run EXPLAIN, read the plan, add/adjust an index, re-run), a schema-design checklist, and one or two executable SQL examples.
Move any retained domain catalogs (engine-specific syntax, index strategy details) into a references/ file and link them one level deep from SKILL.md.
Cut the duplicated 'Purpose' section (it restates the frontmatter description) and keep the 'Use/Do not use this skill when' lists, which are the most useful part of the body.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Roughly 130 of 172 lines are capability catalogs ("Cloud-native databases: Amazon Aurora, Google Cloud SQL...", "Knowledge Base", "Behavioral Traits") that restate what Claude already knows about databases — several heavily padded sections with no new information, matching 'noticeably verbose'. | 2 / 5 |
Actionability | The Instructions offer only high-level hints ("Inspect schema, statistics, and access paths", "Optimize queries and validate with EXPLAIN") with no code, commands, or specific steps; the Capabilities sections describe rather than instruct, matching 'minimal concrete guidance'. | 2 / 5 |
Workflow Clarity | The four Instructions steps are sequenced and mention validation ("Verify correctness and performance under load"), but there is no feedback loop for error recovery, and database operations without explicit validation feedback loops are capped at 3 per the scoring notes. | 3 / 5 |
Progressive Disclosure | Section headers organize the body, but the large capability catalog that belongs in reference files (or nowhere) is fully inlined, and the bundle contains no references/, scripts/, or assets/ files — matching 'some structure but content that should be separate is inline'. | 3 / 5 |
Total | 10 / 20 Passed |