Safe Config Workflow

提供安全修改 OpenClaw 配置的全流程管理，确保变更前确认、自动修复、备份对比及运行验证，保障系统稳定。

This skill appears coherent and limited to local OpenClaw configuration tasks. Before using it: (1) ensure you have a recent backup of ~/.openclaw/openclaw.json (the SKILL already recommends this); (2) review and approve any proposed changes when the skill prompts for confirmation (doctor --fix can automatically change files); (3) be aware tests that intentionally break the Gateway will cause temporary downtime; (4) inspect MEMORY.md if you want to control what lessons are stored locally. If you want extra caution, run the commands manually the first few times instead of letting the agent execute them autonomously.

Type: OpenClaw Skill Name: safe-config-workflow-zh-cn Version: 1.0.1 The skill is designed for 'Safe Config Workflow' and instructs the AI agent to execute various shell commands (`openclaw doctor --fix`, `diff`, `grep -E "..."`, `openclaw gateway status`, `openclaw logs --follow`, `nano`) to manage OpenClaw configuration files. While these commands are plausible and necessary for the stated purpose of configuration management, the direct instruction for the AI to execute shell commands, including a `grep` pipeline for filtering output, represents a powerful capability. This capability, detailed in `SKILL.md`, poses an inherent risk of shell injection or privilege escalation if user input were ever unsafely incorporated into these commands, even though the current implementation uses hardcoded arguments. There is no evidence of intentional malicious behavior like data exfiltration or persistence, but the broad shell execution and file modification permissions warrant a 'suspicious' classification due to the potential for vulnerability.