Kimi Cli Headless Execution

教 OpenClaw 在自动化脚本、后台任务、CI/CD、无 TTY 环境等非交互场景中， 正确使用 Kimi Code CLI 的 `-p/--prompt`、`--print`、`--quiet`、`--wire` 等参数完成无头执行。 涵盖命令构造、自动审批控制、输出格式选择、会话管理、错误处理与安全护栏。

This skill is a coherent how-to for running the 'kimi' CLI in headless/CI contexts, but the manifest fails to declare that it expects the 'kimi' binary and may read local config (e.g., ~/.kimi/config.toml) and use API keys. Before installing or enabling: 1) verify the skill source or prefer a skill with a homepage/repo; 2) treat any run that uses '--yolo' as high-risk—require explicit user consent and avoid running on production or broad directories like $HOME or /etc; 3) run tests in an isolated/sandbox environment first; 4) ensure the agent does not have access to sensitive config files or credentials unless you intentionally grant them; 5) consider using direct API access instead of delegating file-modifying actions to a CLI in automated flows. If the publisher can update the manifest to declare 'kimi' as a required binary and list any config paths or env vars it will access, that would reduce the current inconsistency.

Type: OpenClaw Skill Name: test-rate-limit Version: 0.0.1 The skill bundle (SKILL.md) provides instructions for an AI agent to use the 'kimi' CLI tool in headless mode, specifically encouraging the use of the '--yolo' flag to automatically approve file modifications and shell command executions. While the documentation includes safety guardrails and risk warnings, the core functionality enables high-risk automated system interactions that could be exploited if the agent receives malicious prompts. The discrepancy between the metadata slug 'test-rate-limit' and the actual content 'kimi_cli_headless_execution' is also noted.