Content
75%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.
A well-structured, operational skill body with concrete procedures, validation checkpoints, and good external referencing. The main gap is the absence of explicit error-recovery feedback loops around its destructive database operations.
Suggestions
Add an explicit feedback loop to the destructive procedures: after 'supabase db reset' / 'supabase test db', state 'if reset or tests fail, fix the consolidated SQL and re-run' to reach full workflow clarity.
Tighten the repeated anti-patterns lists across the three subcommands (e.g., consolidate shared schema-vs-migration direction guidance) to reduce token redundancy.
Consider extracting the detailed RLS inversion doctrine and persona matrix into a references file, keeping SKILL.md as the overview that points to it.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Dense and operational, assuming Claude's knowledge of SQL/RLS/supabase with no padding on basics; minor repetition across the per-command anti-pattern lists could be trimmed. | 4 / 5 |
Actionability | Numbered procedures with concrete commands ('supabase db reset', 'supabase test db'), real migration filenames, idempotent SQL forms, and a classification table give mostly executable guidance with minor gaps. | 4 / 5 |
Workflow Clarity | Each subcommand has a clear sequenced procedure with validation checkpoints (verify-with-user, db reset, test db); the destructive-operation cap does not apply since validation is present, but explicit validate-fix-retry loops are not spelled out. | 4 / 5 |
Progressive Disclosure | Well-organized sections with clearly signaled one-level references to supabase/AGENTS.md (with anchors) and opt-library rather than duplicating; no bundle files exist, and some inlined RLS doctrine could arguably be split out. | 4 / 5 |
Total | 16 / 20 Passed |