使用方法
- 复制安装命令，在自己的智能体环境安装前先检查源码包。
- 从原始文档核对所需工具、凭据和支持的宿主。
- 说明输入材料与期望交付物，并对照来源检查输出。
npx skills add https://github.com/robsonrung/rar-skills --skill grok-runner开始前需了解
- 安装不包含模型订阅、API 额度或访问文件的权限。
- 仓库星数和安装量反映活跃情况，不等于独立质量或安全测试。
按源仓库内容呈现，保留标题、案例、代码、表格、链接以及原文引用的演示图片。
Grok Runner
Execute the caller's scoped prompt through the local CLI. Shared roles, envelope keys, and result handling live in shared/references/runner-common.md. Preserve seat fidelity: name the provider that actually answered, and report unavailable or unverified seats honestly.
Invocation
Set SKILL_DIR to this loaded skill directory in the same shell call:
SKILL_DIR="<absolute path of this skill directory>";
python3 "$SKILL_DIR/scripts/run_grok.py" "<scoped prompt>" --jsonAn approved workflow supplies the exact model, effort, role, tool mode, and receipt policy. Pass those approved options and --disable-fallback; this basic example does not authorize a different route.
Runtime boundaries
- Use the local authenticated
grokCLI in print mode. Analysis roles default to enforced read-only plan mode. --restrict-toolsforces plan mode;--allow-writeopts out within the caller's authority. Prompt and file context may be sent to the configured provider.- Sessions persist under
~/.grok; the native CLI has no persistence opt-out. - The native
modelUsagereceipt identifies the serving model. Effort values unsupported by a direct call are clamped and reported; approved routes must satisfy their exact plan. - No fallback is allowed. A missing CLI is an unavailable seat, never another model's answer.
- Structured output is checked locally after the native schema constraint. A zero exit code does not make malformed output successful.
Load by need
- references/command-reference.md: flags, examples, native mappings, and return codes when composing a non-basic invocation.
- references/runtime-reference.md: authentication, fallback conditions, configuration, and extended envelope fields when configuring a route or interpreting its result.
- references/continuation.md: only when resuming a session or managing a background job.
Result
Read agent_message and the wrapper envelope. Check success, serving-model receipt, effective runner/provider, and any fallback or malformed-output status. A completed process alone does not prove the requested task or model was used. Report touched files when the caller authorized edits.
Integration
agents/openai.yaml exposes this skill as a native Codex-app subagent seat; do not remove it.
