system-repair-expert

提供严格按优先级步骤诊断与解决系统、软件、配置及报错问题，结合记忆匹配、官方文档和多来源验证确保方案高效可靠。

This skill appears coherent for a troubleshooting assistant, but check these before installing: - Confirm how your Clawd platform provides web_search (Brave) and memory services and whether any external API keys are managed by the platform (the skill itself doesn't request keys). - Confirm whether the skill will persistently store case data in your memory system; SKILL.md says it will store/update memory but the manifest/code don't clearly show memory-write—ask the author or platform what memory permissions will be granted. - The skill promises to ask for explicit consent before generating or running any repair scripts; be sure the agent prompts you and you never allow script execution without reviewing the script and rollback steps. - Avoid sending sensitive secrets (passwords, private keys) in logs—provide only the information necessary for debugging. - If you plan to publish any created skills to a hub, review them for accidental exposure of credentials or hard-coded endpoints. If you want, I can extract the exact lines that mention memory-write and the lines in the manifest/package.json that differ so you can ask the author specific questions.

Type: OpenClaw Skill Name: system-repair-expert Version: 1.0.0 The skill is classified as suspicious due to its inherent high-risk capabilities, specifically the instructions in `SKILL.md` for the AI agent to '调用内置的「创建/编写 Skill」能力' (call the built-in 'create/write Skill' capability) and '编写一次性修复脚本/命令' (write one-time fix scripts/commands). While these capabilities are presented with explicit safety guardrails, user consent requirements, and are framed as a last resort for legitimate system repair, they represent powerful actions that, if misused or exploited, could lead to unauthorized code execution or skill creation. The `CLAWDHUB_MANIFEST.json` also declares `file_system_read` capability, which, while plausible for a repair expert, adds to the overall risk profile.