Content
92%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.
A tight, well-structured, instruction-only skill: it specifies the exact command, pre-flight validation, failure conditions, and safe permission/authentication policies with almost no wasted tokens. The only notable gap is that the JSON event stream the skill requires Claude to parse is never exemplified.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The body is lean and operational throughout — "set a finite timeout, normally 10 minutes", "Do not add `--auto` merely to avoid a denied permission" — with zero concept explanations or padding; every token earns its place. | 5 / 5 |
Actionability | Gives the copy-paste-ready command "opencode run \"<prompt>\" --format json" and the pre-flight check "command -v opencode", but instructs Claude to "Parse stdout as a JSON event stream and treat an error event ... as failure" without showing the event schema, a minor gap. | 4 / 5 |
Workflow Clarity | Clear numbered sequence with a pre-flight validation step ("Check availability with command -v opencode"), explicit failure conditions ("error event or nonzero exit"), and defined fallback behavior ("stop and ask the user"); the single action is unambiguous, and the skill is not destructive or batch (workspace changes are explicitly gated), so no validation cap applies. | 5 / 5 |
Progressive Disclosure | The skill is under 50 lines, self-contained with no bundle files, and organized into well-signaled sections ("Run", "Authentication And Permissions"), which per the rubric's simple-skill guideline earns full credit. | 5 / 5 |
Total | 19 / 20 Passed |