Content
72%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 highly actionable with executable code and commands across five tools, but its multi-step destructive/batch workflows lack explicit validation feedback loops, and it is a monolithic single file with no progressive disclosure into reference files.
Suggestions
Add explicit validation/verification feedback loops to destructive and batch workflows (e.g., after a batched backfill, verify row counts and confirm no NULLs remain before proceeding; after a migration, run a verification query before declaring success).
Consolidate the redundant 'When to Activate' and 'When to Use This Skill' lists into a single section to reduce duplication.
Move the per-tool reference material (full Prisma/Django/Drizzle/golang-migrate sections) into separate reference files under references/ and keep SKILL.md as an overview with one-level-deep links.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Mostly lean with terse GOOD/BAD blocks, a checklist, and an anti-patterns table, but the 'When to Activate' and 'When to Use This Skill' lists overlap and the expand-contract pattern is described twice, leaving minor tightening possible. | 4 / 5 |
Actionability | Provides copy-paste-ready executable SQL (CONCURRENTLY indexes, batched DO$$ backfill), real CLI commands for all four ORMs, and real schema/migration code covering the common cases. | 5 / 5 |
Workflow Clarity | A pre-apply safety checklist and expand-contract timeline give a rough sequence, but for destructive/batch database operations the rubric caps workflow_clarity at 3 when explicit validate->fix->retry feedback loops are absent, and no post-migration verification step is provided. | 3 / 5 |
Progressive Disclosure | No bundle files exist and the 337-line body is a single monolithic file with no references to separate per-tool reference files; section headers provide structure but content that could be split (full Prisma/Django/Drizzle sections) is inlined with no file-level navigation. | 3 / 5 |
Total | 15 / 20 Passed |