Content
46%Scale 1-5Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The skill provides comprehensive coverage of Supabase operations via Rube MCP with well-organized workflow sections and useful pitfall documentation. However, it suffers from significant verbosity through repeated pitfall information, over-documentation of parameters Claude could discover via tool schemas, and a monolithic structure that would benefit from splitting into overview + reference files. It also lacks validation/feedback loops for destructive database operations.
Suggestions
Split the detailed parameter documentation, pitfalls catalog, and quick-reference table into a separate REFERENCE.md file, keeping SKILL.md as a concise overview with tool sequences and key gotchas only.
Remove the duplicated pitfalls — consolidate into the 'Known Pitfalls' section only and remove per-workflow pitfall lists, or vice versa.
Add explicit validation steps to the database query workflow: after INSERT/UPDATE/DELETE via RUN_SQL_QUERY, include a step to verify the change (e.g., 'Run a SELECT to confirm the mutation succeeded') with error recovery guidance.
Remove parameter details that Claude can discover via RUBE_SEARCH_TOOLS (as the skill itself instructs to 'always call RUBE_SEARCH_TOOLS first to get current tool schemas'), keeping only non-obvious gotchas and format requirements.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The content is noticeably verbose with significant repetition — pitfalls are listed in each workflow section and then repeated again in a consolidated 'Known Pitfalls' section. Many parameter details (like PostgREST filter operators) are things Claude already knows or could discover via RUBE_SEARCH_TOOLS. The document could be cut by 40-50% without losing actionable information. | 2 / 5 |
Actionability | The skill provides concrete tool names, specific parameter names and formats, and clear tool sequences for each workflow. However, there are no executable code/command examples showing actual tool invocations with sample payloads, which would make it fully copy-paste actionable. The guidance is specific enough to act on but lacks worked examples. | 4 / 5 |
Workflow Clarity | Tool sequences are clearly numbered with prerequisite/required/optional annotations, which is good. However, for database operations (which include destructive SQL via RUN_SQL_QUERY), there are no explicit validation checkpoints or feedback loops — no 'verify the result after INSERT/UPDATE/DELETE' steps. Per the rubric, missing feedback loops for database operations caps this at 3. | 3 / 5 |
Progressive Disclosure | This is a monolithic document with no bundle files and no references to external files. At 200+ lines with detailed parameter references, pitfalls sections, and a full quick-reference table, much of this content (e.g., the complete parameter documentation, the pitfalls catalog, the quick reference table) would be better split into separate reference files. Everything is inlined in one large file. | 2 / 5 |
Total | 11 / 20 Passed |