Content
85%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
A highly actionable, well-structured skill body with clear sequencing and verification checkpoints; the only drag is mild verbosity from the redundant intro and the inline root-resolution block.
Suggestions
Drop the Japanese intro sentence (or trim it to a one-line note) since the description already covers purpose, removing redundant tokens.
Consider moving the HARNESS_PLUGIN_ROOT resolution into a referenced helper script so the Flow section stays focused on the per-flag commands, with a single source call instead of the inline probe loop.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is mostly lean executable bash, but the Japanese intro line restates the description and the inline HARNESS_PLUGIN_ROOT resolution block is long; it could be tightened or delegated to a referenced helper. | 2 / 3 |
Actionability | Each flag branch gives fully executable, copy-paste-ready commands, including concrete cursor-agent binary detection and specific script invocations. | 3 / 3 |
Workflow Clarity | The per-flag sequence is explicit, with --show verification after mutations, --check validation, and error-exit codes providing validation checkpoints. | 3 / 3 |
Progressive Disclosure | It is a single-purpose, well-organized SKILL.md (Quick Reference, Rules, Flow, Output) with no nested reference chains and no need for separate bundle files. | 3 / 3 |
Total | 11 / 12 Passed |