Content
81%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 strong, dense, prescriptive body: clear stepped workflow with per-step completion criteria, concrete validation commands, and specific tool patterns. Its only weaknesses are minor — a redundant recap section and detail sections that could be split into reference files.
Suggestions
Trim the 'Final self-check' checklist or fold it into the Steps' completion criteria, since it restates rules already covered section by section
Move the detailed 'Fixtures and cleanup' and 'Mocking' rule lists into a references/ file, keeping SKILL.md as a lean overview with one-level-deep pointers
Add one complete example test (Arrange-Act-Assert with a fixture and a mocker.patch.object call assertion) to make the common case fully copy-paste ready
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is entirely prescriptive rules with no teaching of known concepts, and rationales are single-clause ('an earlier failure skips them'). The 'Final self-check' section recapitulates rules already stated in earlier sections, which is minor redundancy keeping it below the lean anchor at 5. | 4 / 5 |
Actionability | Concrete, executable commands are given ('uv run pytest dimos/path/to/test_file.py -k test_name -m \'not (self_hosted or self_hosted_large)\'', './bin/pytest-fast', 'pre-commit run --all-files') along with precise patterns ('patch the target with mocker.patch.object(...), run the behavior, then assert with assert_called_once_with(...)'). Minor gap: no complete example test snippet, so it is 'mostly executable' rather than fully copy-paste ready across common cases. | 4 / 5 |
Workflow Clarity | The five numbered steps each carry an explicit 'Completion:' validation checkpoint, followed by concrete validation commands and a final self-check checklist. This matches the anchor for a clear sequence with explicit validation steps and a checklist; unit-test edits are not destructive/batch operations, so no cap applies. | 5 / 5 |
Progressive Disclosure | No bundle files exist, so this scores the body structure itself: sections are well-organized and the external doc pointers are clearly signaled and one level deep ('Consult these only when the branch needs more context'). At ~95 lines with detail-heavy sections (fixtures/mocking rules) inlined in a single file, it stops short of the clean overview-plus-split-content anchor at 5. | 4 / 5 |
Total | 17 / 20 Passed |