Content
87%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The body is an efficient, fully executable CLI reference with clean sectioning and only need-to-know notes. Its one weakness is workflow clarity: batch and destructive commands lack explicit validation/verification checkpoints.
Suggestions
Add a verification checkpoint before batch/destructive runs: e.g., run `clawhub list` to review installed skills before `clawhub update --all --force --no-input`, since --no-input skips confirmation.
Make the publish flow an explicit auth-then-verify-then-publish sequence: `clawhub login` → `clawhub whoami` (confirm identity) → `clawhub publish ...`, and note to double-check slug/version since publishing is hard to reverse.
Close the update feedback loop by noting how to confirm success, e.g. re-run `clawhub list` after `clawhub update` to verify the new version is installed.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is a lean command reference grouped by task with almost no prose, and the Notes bullets cover only non-obvious behavior (hash-based update matching, registry/workdir overrides). It is not a 2 because there is no unnecessary explanation of concepts Claude already knows. | 3 / 3 |
Actionability | Every section gives fully executable, copy-paste-ready commands with real flags, e.g. `clawhub install my-skill --version 1.2.3` and `clawhub publish ./my-skill --slug ... --version ... --changelog ...`. It is not a 2 because the examples are complete and concrete rather than pseudocode or missing key details. | 3 / 3 |
Workflow Clarity | Commands are clearly grouped by operation, but batch/destructive invocations like `clawhub update --all --force --no-input` and `clawhub publish` are shown without explicit validation or verification checkpoints, which the rubric caps at 2. It is not a 3 because the cap for missing validation on batch operations applies; it is not a 1 because steps are present and organized rather than unclear. | 2 / 3 |
Progressive Disclosure | At roughly 50 lines with no bundle files and no need for external references, the well-organized Install/Auth/Search/Update/List/Publish/Notes sections satisfy the simple-skill allowance for a top score. It is not a 2 because content is appropriately sectioned inline with no nested or poorly-signaled references. | 3 / 3 |
Total | 11 / 12 Passed |