Content
62%Reviews 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-sequenced, validation-aware instruction spec with concrete field/value constraints, but it is somewhat verbose, lacks a full worked example, and keeps all detail inline rather than splitting it across reference files.
Suggestions
Add one complete candidate JSON example showing the exact reply shape and required fields.
Move the detailed root/selection/integration/policy field schema into a reference file and link to it from the body.
De-duplicate the repeated 'do not' enumeration and verification-refs wording to tighten token usage.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | It assumes Claude's competence and avoids re-explaining basics, but the ~130 lines repeat several 'do not' enumerations and restate the verification-refs requirement, and could be tightened. | 2 / 3 |
Actionability | Field names and literal values (e.g. 'partial_or_blocked', 'replan_required', the schema id, ID rules) are highly specific, yet no complete worked JSON candidate example is provided to make it copy-paste ready. | 2 / 3 |
Workflow Clarity | It gives explicit ordering ('Apply these selection rules before emitting the candidate'), validation checkpoints requiring direct argv verification commands, and a replan_required feedback loop for error recovery. | 3 / 3 |
Progressive Disclosure | No bundle files exist and all content sits inline in one >50-line file; the detailed field/schema spec could be split into a reference file with clear navigation. | 2 / 3 |
Total | 9 / 12 Passed |