Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
Ai Company Cpo
v2.1.0AI公司首席公共官（CPO）技能包。企业信誉资产守护者、品牌声誉建设、分层媒体网络、四级危机预警、黄金4小时响应、AI舆情监测。
⭐ 0· 126·1 current·1 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Suspicious
medium confidencePurpose & Capability
The skill is clearly a public-affairs / crisis-response playbook and templates (monitoring, response timelines, media segmentation). Requesting network API access and file read/write can be justified for monitoring, report generation, and evidence storage, but the SKILL.md does not tie these permissions to specific services or explain why higher-risk capabilities (see mcp) are required.
Instruction Scope
SKILL.md content is primarily procedural guidance, role definitions, and interface schemas; it does not instruct the agent to read arbitrary system files or exfiltrate secrets. However it references automation features ('自动化申诉/一键提交', '区块链存证', '生成式搜索引擎中植入') without specifying endpoints, auth flows, or which files it will read/write — leaving broad discretion to the runtime agent.
Install Mechanism
Instruction-only skill with no install spec and no code files — this minimizes supply-chain risk (nothing is downloaded or written during install).
Credentials
The skill declares no required env vars but requests network (api) and file read/write permissions and depends on other skills. For automated submission to platforms and blockchain anchoring, credentials would typically be needed; the SKILL.md does not declare or justify required credentials, and generic network access plus file I/O could be used to access or transmit sensitive data if misused.
Persistence & Privilege
The permissions include mcp: [sessions_send, subagents], which implies the skill can create or orchestrate subagents and send sessions. That capability is powerful and not explained by the document's playbook-style content. Although always:false limits forced inclusion, autonomous invocation plus subagent creation increases the blast radius if the skill is granted these permissions.
What to consider before installing
This skill appears to be a coherent CPO/PR playbook, but it requests runtime permissions that are broader than the document explains. Before installing: (1) Ask the publisher which external APIs and endpoints the skill will call and require explicit allowlists for those domains; (2) Require the skill to declare any credentials it needs and avoid granting blanket network access until those are scoped; (3) Deny or tightly constrain mcp/subagent capabilities unless you understand and trust automatic subagent creation in your environment; (4) If you must grant file read/write, restrict the accessible paths to a safe workspace; (5) Prefer a version that documents exact integrations (monitoring providers, blockchain service, complaint submission endpoints) or run the skill in a sandboxed environment first. If the publisher cannot justify the mcp/subagent and generic network/file permissions, treat the skill as high risk.Like a lobster shell, security has layers — review code before you run it.
ai-companyvk97e99zhfdze7q77rk83t5y47h84s1ssc-suitevk9742gk4s47a7ahzmpcsj469q584mdcqcpovk97e99zhfdze7q77rk83t5y47h84s1sscsatvk97e99zhfdze7q77rk83t5y47h84s1sslatestvk972kpgnawxygz9mf15rtbwaq984v2x4npsvk97e99zhfdze7q77rk83t5y47h84s1ssprvk979hqd337nxthp3ec4jvvbkfs84m2vmproduct-strategyvk97e99zhfdze7q77rk83t5y47h84s1ssreputationvk979hqd337nxthp3ec4jvvbkfs84m2vmuxvk97e99zhfdze7q77rk83t5y47h84s1ss
License
MIT-0
Free to use, modify, and redistribute. No attribution required.