Content
77%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This is a solid, actionable PostgreSQL skill with excellent executable examples and a well-structured workflow that includes proper validation checkpoints. Its main weaknesses are moderate verbosity (some sections like 'When to Use' and 'Knowledge Reference' add little value) and the absence of the referenced bundle files that the reference table points to. The constraints section is well-organized with clear MUST/MUST NOT boundaries.
Suggestions
Remove or significantly trim the 'When to Use This Skill' section (duplicates the description) and the 'Knowledge Reference' flat list (Claude already knows these terms).
Move the 'Common Patterns' section content into the referenced files (references/performance.md, references/jsonb.md, etc.) to keep SKILL.md as a lean overview and actually provide the bundle files.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is mostly efficient but includes some unnecessary content. The 'When to Use This Skill' section largely duplicates the description, the 'Knowledge Reference' section at the bottom is a flat list of terms Claude already knows, and the 'Output Templates' section is somewhat generic. However, the code examples and constraints are tight and useful. | 2 / 3 |
Actionability | The skill provides fully executable SQL examples throughout — from identifying slow queries via pg_stat_statements, to creating partial indexes with CONCURRENTLY, to JSONB GIN indexing, VACUUM monitoring, and replication lag checks. All code is copy-paste ready with inline comments explaining what to look for. | 3 / 3 |
Workflow Clarity | The end-to-end example provides a clear 5-step sequence (identify → analyze → fix → verify → update stats) with explicit validation checkpoints (verify index is used with EXPLAIN, confirm lower actual time). The core workflow also includes verification steps ('verify with EXPLAIN before deploying', 'verify improvements after each change'). | 3 / 3 |
Progressive Disclosure | The reference table clearly signals five separate reference files with 'Load When' context, which is good structure. However, no bundle files are provided, so the referenced files don't actually exist. Additionally, the SKILL.md itself is fairly long (~120 lines of content) with inline patterns that could arguably live in the referenced files, making the main file heavier than ideal. | 2 / 3 |
Total | 10 / 12 Passed |