Content
92%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.
Excellent instructional content: a clear version-routing decision procedure, complete minimal YAML examples for both spec generations, concrete validation commands with feedback loops, and a well-structured one-level reference bundle. The only tightening opportunities are the introductory component definitions and the generic external-content boilerplate.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is dense and domain-specific (routing rules, minimal YAML examples, validation commands, pitfalls table) with little padding, but a few sections could be trimmed: the opening definition list ('Semantic models - Metadata configurations that define...') restates somewhat obvious context, and 'Handling External Content' is generic boilerplate. This fits anchor 4 (efficient, minor instances of over-explanation) better than anchor 5's 'every token earns its place', and is far above the verbosity of anchor 3. | 4 / 5 |
Actionability | Two complete, copy-paste-ready YAML examples for both spec variants ('# models/fct_orders.yml' and '# models/sem_orders.yml'), exact commands ('dbt parse', 'dbtf parse', 'dbt sl validate', 'mf validate-configs', 'uvx dbt-autofix deprecations --semantic-layer'), and concrete filter syntax templates cover the common cases. This matches anchor 5 (fully executable, copy-paste ready, covers common cases); anchor 4 would imply missing key details, which are not evident. | 5 / 5 |
Workflow Clarity | The spec-determination workflow is explicitly sequenced (Step 1: check existing config → Step 2: route by spec and dbt version → Step 3: follow the guide) and validation is enforced with feedback loops: two-stage validation, 'Do not consider work complete until both validations pass', and the stale-manifest recovery instruction to re-run 'dbt parse' before 'mf validate-configs'. This matches anchor 5 (clear sequence, explicit validation, error-recovery guidance); anchor 4 would require missing checkpoints, which are present. | 5 / 5 |
Progressive Disclosure | SKILL.md functions as an overview with well-signaled, one-level-deep references: 'Additional Resources' lists all four files (time-spine, best-practices, latest-spec, legacy-spec), each of which exists in references/ and is self-contained (no nested reference chains). Minimal examples are inline while full YAML specs are appropriately split out, matching anchor 5. | 5 / 5 |
Total | 19 / 20 Passed |