Content
73%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 well-structured, highly actionable runbook: exact commands, explicit input validation with failure handling, background execution, concrete result-presentation and error-recovery guidance. Its main weaknesses are the verbatim duplication of the command blocks between steps 2 and 3 and a slightly incorrect Task-tool invocation instruction.
Suggestions
Deduplicate the command blocks: state the full commands once in '### 2. Build the command' and have '### 3. Run the review in the background' reference them ('launch the command from step 2 as a background task'), cutting ~16 lines.
Correct the background-launch instruction to match the harness (Bash tool with run_in_background rather than 'Task tool with subagent_type: "Bash"') so the launch step is executable as written.
Trim the bullets under the step-2 command ('If --base is specified, include it...') that restate points already made in step 1 and the surrounding prose.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is generally lean and assumes competence, but the full bash command blocks are duplicated verbatim in '### 2. Build the command' and '### 3. Run the review in the background' (roughly 16 repeated lines), and the bullets after the step-2 commands restate what the surrounding prose already says. This is more than the 'minor instances' of anchor 4, fitting 'mostly efficient but could be tightened'. | 3 / 5 |
Actionability | The 'roborev review --from-skill --branch --wait --type design' and 'git rev-parse --verify --end-of-options' commands are copy-paste ready, with concrete error remedies ('roborev daemon status', 'roborev init') and worked example dialogues. The one gap is 'Use the Task tool with run_in_background: true and subagent_type: "Bash"', which does not match the harness's actual background-bash mechanism — concrete guidance with a minor executability gap (anchor 4). | 4 / 5 |
Workflow Clarity | A clear five-step sequence (validate inputs, build command, run in background, present results, offer next steps) with an explicit validation checkpoint — 'git rev-parse --verify --end-of-options' plus 'If validation fails... Do not proceed' — and error-recovery guidance for daemon/init failures. This is not a destructive or batch operation, so no cap applies, and the checkpoints and feedback loops match anchor 5. | 5 / 5 |
Progressive Disclosure | A single-file skill with no bundle directories; all content (commands, workflow, panels, examples, see-also) appropriately lives in SKILL.md under clearly headed, navigable sections. Minor organization gaps — the duplicated sections 2/3 and an Examples section that could be trimmed — keep it at anchor 4 rather than 5. | 4 / 5 |
Total | 16 / 20 Passed |