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 body is a compact, well-organized index with a genuinely useful quick-reference section, but it functions as a pointer document whose targets are missing: all 14 referenced rule files are absent from the bundle, leaving the skill with little standalone actionable content and no workflow guidance. Inlining the key rules or shipping the rule files with descriptive labels would resolve most weaknesses.
Suggestions
Ship the 14 rule files referenced under 'Rule Files' — or inline the most important rules into the body — since every referenced path is currently dangling.
Annotate each rule-file entry with what it covers (e.g., 'rules/endpoint-optimization.md — query tuning for API endpoints') and map the 'When to Apply' bullets to the relevant file so readers know which file to open for which task.
Add a minimal workflow with a validation checkpoint (e.g., edit datafile -> run tb build/tests locally -> tb deploy) and one or two executable tb command examples to raise actionability and workflow clarity.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean with no concept explanations Claude already knows — sections are 'When to Apply', a rule-file index, and a dense 'Quick Reference' of genuinely non-obvious facts ('Build target comes from tinybird.config.json dev_mode', 'Use MergeTree by default; AggregatingMergeTree for materialized targets'). It falls short of 5 only because the intro line ('Guidance for Tinybird file formats, SQL rules, optimization patterns, and data modeling...') duplicates the frontmatter description and overlaps the 'When to Apply' list. | 4 / 5 |
Actionability | The Quick Reference contains some concrete, specific guidance (named commands like 'tb sql and tb logs default to local unless --cloud or --branch=<branch-name>', config keys, engine choices), but the actual executable detail is deferred to the 14 rule files and no usage examples or command invocations appear inline — matching 'Some concrete guidance but incomplete... missing key details' rather than score 4's 'concrete code or commands with minor gaps'. | 3 / 5 |
Workflow Clarity | There is no multi-step sequence anywhere: 'When to Apply' is a list of situations, not ordered steps, and no validation/checkpoint guidance exists (deploy and batch operations are in scope yet unvalidated). It also doesn't qualify for the simple-skill exception because it is an index into rule files with no mapping of which task reads which file, so the reader cannot form a clear procedure — matching 'Steps listed but validation gaps; sequence present but checkpoints missing' at best. | 3 / 5 |
Progressive Disclosure | The body is structured as a one-level index of 14 rule files ('rules/datasource-files.md', 'rules/pipe-files.md', ...), but none of those files exist in the bundle — no rules/, references/, scripts/, or assets/ directories are present, so every reference is dangling and navigation dead-ends. The list entries are also bare paths with no per-file description of what each covers, fitting 'Minimal structure... references are buried [or broken]' rather than score 3's 'references present but not clearly signaled'. | 2 / 5 |
Total | 12 / 20 Passed |