Content
57%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The body is lean and organized with a clear verification stance, but it lacks a concrete sequenced workflow and executable specifics, relying on vague placeholders like '(research/command)'. It is a reasonable skeleton that would benefit from sharper actionability.
Suggestions
Convert the 'You must' bullet list into a numbered, ordered workflow so the sequence (research -> ask -> verify -> proceed) is explicit.
Replace the vague '(research/command)' with a concrete example of how to spawn and run a verification task.
Drop the advisory tone-setting sentence and merge the redundant verification bullets to improve token efficiency.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is short and avoids explaining concepts Claude already knows, but the advisory line ('skeptical, thorough...high-quality technical specifications') is mild padding and the two verification bullets ('Spawn new research tasks to verify' and 'Only proceed once you've verified the facts yourself') overlap and could be tightened. | 2 / 3 |
Actionability | It gives a concrete output path ('.cursor/plans/{plan_name_folder} ... .md') and a verification directive, but '(research/command)' is vague and there are no specific commands or examples, leaving key details incomplete. | 2 / 3 |
Workflow Clarity | There is an explicit verify-before-proceed checkpoint ('Only proceed once you've verified the facts yourself'), but the steps are an unordered obligation list rather than a clearly sequenced workflow, and the checkpoints are otherwise implicit. | 2 / 3 |
Progressive Disclosure | This is a simple skill under 50 lines with no external bundle files; the content is organized into a heading and a bulleted must-do list, which the rubric's simple-skills note permits to score 3. | 3 / 3 |
Total | 9 / 12 Passed |