Content
50%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The skill covers TiDB test placement and writing conventions comprehensively with a reasonable workflow and useful package-specific guidance. However, it lacks executable code examples (e.g., a sample test function skeleton), validation checkpoints in the workflow, and the package-specific notes section is dense enough that it would benefit from being extracted into a reference file. The content assumes Claude's competence appropriately but could be more actionable with concrete examples.
Suggestions
Add a concrete, executable example of a properly structured TiDB test function (with testkit, store setup, table-driven cases, and assertions) to improve actionability.
Add explicit validation steps to the workflow, such as 'Run `make bazel_test` to verify no regressions' after moving tests or updating BUILD.bazel.
Extract the lengthy 'Package-specific notes' section into a separate reference file (e.g., `references/package-test-notes.md`) and link to it from the SKILL.md to improve progressive disclosure and conciseness.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The content is mostly efficient and avoids explaining basic concepts, but several bullet points in 'Package-specific notes' are quite dense and could be tightened. Some rules are overly specific (e.g., the explain-format paragraph) and border on verbose for a skill overview. | 2 / 3 |
Actionability | The skill provides concrete naming conventions, variable names, and specific commands (e.g., `rg --files -g '*_test.go'`, `-record -tags=intest,deadlock`), but lacks executable code examples showing a complete test structure. Many instructions are directive but not copy-paste ready—no example test function is provided. | 2 / 3 |
Workflow Clarity | The 'Quick workflow' section provides a clear 5-step sequence, but lacks explicit validation checkpoints or feedback loops. For operations like moving tests and updating BUILD.bazel (which can break builds), there's no 'verify build passes' or 'run affected tests' step before proceeding. | 2 / 3 |
Progressive Disclosure | References to external files like `references/planner-guide.md`, `references/<pkg>-case-map.md`, and `docs/agents/testing-flow.md` are present and clearly signaled. However, no bundle files were provided to verify these exist, and the package-specific notes section is quite long and could benefit from being split into a separate reference file rather than being inline. | 2 / 3 |
Total | 8 / 12 Passed |