Content
53%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 content provides substantial executable, ORM-specific migration code with run/rollback commands, making it reasonably actionable. However, it is hurt by repeated verbose patterns, the absence of validation checkpoints for destructive migrations, and a Resources section that references bundle files which do not exist.
Suggestions
Move the bulk code examples into the referenced files (references/schema-migration.md, references/rollback-strategies.md, etc.) and keep SKILL.md as a concise overview — and ensure those files actually exist, fixing the inconsistent "resources/" vs "references/" path on line 23.
Add explicit validation checkpoints to destructive/batch workflows (e.g. run migrations on staging, verify row counts/backfill completeness, then proceed), with a validate→fix→retry feedback loop, to lift workflow clarity above 3.
Tighten repeated Sequelize up/down boilerplate and remove narration of basic steps ("Get all records", "Copy and transform data") that Claude already infers from the code.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is mostly executable code (efficient) but repeats near-identical Sequelize up/down patterns across many sections and includes redundant narration ("// Get all records", "// Copy and transform data", "// Handle differences") that Claude does not need. It is mostly efficient with some tightening possible, matching the 3 anchor rather than the lean 5. | 3 / 5 |
Actionability | Most sections give executable code with explicit run/rollback commands (e.g. "npx sequelize-cli db:migrate", "npm run typeorm migration:run", "npx prisma migrate deploy"), covering common cases well. It is not a 5 because Prisma rollback/down is omitted and a few examples (cross-database, checkpoint rollback) are partial. | 4 / 5 |
Workflow Clarity | This is a destructive/batch database-migration skill, yet the multi-step sections (renaming columns, type changes, blue-green deploys) lack explicit validation checkpoints or feedback loops; the rubric caps such skills at 3. It is above 2 because sequences are present, but the missing validate-then-proceed gating prevents a higher score. | 3 / 5 |
Progressive Disclosure | The body inlines ~400 lines of examples that clearly belong in reference files, and the Resources section lists files (references/*.md, assets/*, scripts/*) that do not exist in the skill directory; line 23 also references a non-existent "resources/implementation-playbook.md" with an inconsistent path prefix. Broken/missing bundle files plus inlined bulk content match the 2 anchor (minimal structure, references buried/invalid). | 2 / 5 |
Total | 12 / 20 Passed |