Content
88%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 exceptionally actionable skill: executable test code for every pattern, explicit sequencing with validation checkpoints and a failure-diagnosis feedback loop, and clearly signaled references. The minor costs are a slightly long body with a redundant trailing References section and second-level navigation to the per-engine reference files.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense with non-obvious, load-bearing specifics (the RLS connection-role matrix, "Foreign key constraint checks, Unique constraint checks... bypass RLS", the PgBouncer transaction-pooling caveat) with no basic-concept padding. Minor trims exist — the References section duplicates links already given inline (the OWASP URL appears twice) and Test 8's stacked comments are slightly chatty — so it is efficient but not perfectly lean. | 4 / 5 |
Actionability | Fully executable guidance throughout: complete copy-paste Python tests for all eleven patterns, a verification query with expected output ("Expect: rolsuper=f, rolbypassrls=f"), a working tenant_tx context manager, concrete policy DDL, and framework-specific config locations for Django, Rails, and Spring Boot. | 5 / 5 |
Workflow Clarity | Sequencing is explicit ("Plan first, then implement" → Steps 1–4), Step 1 embeds a validation checkpoint, and Step 4 is a genuine error-recovery feedback loop ending with "Write the minimal regression test before fixing"; the anti-patterns table doubles as a checklist. Validation is present for risky operations, so the workflow cap does not apply. | 5 / 5 |
Progressive Disclosure | References are well-signaled inline at point of use ("ready-made skeletons are in [references/framework-skeletons.md]", "The full workflow is in [references/ci-workflow.md]") and all referenced files exist, one primary level deep. However, the ~470-line body inlines the entire eleven-test battery, and several bundle files (mysql-mariadb.md, sqlserver.md, vitess.md, cockroachdb.md, models.md, test-surfaces.md) are reachable only second-level through row-level-security.md or other-engines.md rather than from SKILL.md itself — good structure with minor organization gaps. | 4 / 5 |
Total | 18 / 20 Passed |