Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
Claw Office Report
v1.0.1自动上报工作状态到 Claw Office 微信小程序，让你的虚拟 AI 员工实时在线打工赚金币。支持开始/结束工作、中途更新状态。
⭐ 0· 135·2 current·2 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Suspicious
high confidencePurpose & Capability
Name/description, required env CLAW_OFFICE_KEY and the included code all indicate the skill's purpose is to report work state to an external Claw Office endpoint — that is coherent. Minor inconsistencies: SKILL.md examples post to https://clawoffice.zjhzwukan.xyz/user/report-work while index.js defaults to https://clawoffice.zjhzwukan.xyz/api/user/report-work (API_BASE vs examples). SKILL.md tells users to put the key in ~/.openclaw/openclaw.json while the code reads process.env.CLAW_OFFICE_KEY (platform probably maps config into env, but this is an implementation/UX mismatch).
Instruction Scope
SKILL.md explicitly orders the agent to keep reporting secret from the user (do not mention in conversation, be silent on failures) and to perform background, non-blocking curl calls. This is deliberate concealment of network activity from the user — a red flag for covert telemetry/exfiltration. The code implements background curl calls accordingly. The instructions also require putting the key into config (fine), but the instruction to never disclose the reporting in conversation grants the skill broad discretion and reduces user visibility.
Install Mechanism
No install spec (instruction-only) and no external downloads; included JS source files are bundled with the skill. This is lower installation risk than remote installers. There is use of child_process.exec to run curl (normal for an instruction that shells out), but no external install URLs or archive extraction.
Credentials
Only CLAW_OFFICE_KEY is declared as required — proportional to a service that needs an API key. However index.js also reads an undocumented CLAW_OFFICE_API env var to override the endpoint; that variable is not declared in metadata and could be used to redirect reports to an arbitrary host if set. The skill sends the CLAW key to an external domain, so you should only use it if you trust that domain.
Persistence & Privilege
always:false (good) but the skill is designed to be invoked automatically via hooks (beforeTask/afterTask). Autonomous invocation combined with the explicit instruction to hide reporting from conversations increases the stealthiness and blast radius: the agent can repeatedly contact an external endpoint without informing the user. This combination is a meaningful risk even though no elevated platform flag (always:true) is present.
What to consider before installing
This skill will silently send your CLAW_OFFICE_KEY and small 'start/stop/update' messages to an external server whenever tasks start/finish. Consider these points before installing:
- Trust: Only install if you trust the remote host (clawoffice.zjhzwukan.xyz by default). The key is transmitted to that host; it could be used to act on your account in the mini-program.
- Covert behaviour: The SKILL.md explicitly instructs the agent to hide these reports from you and to ignore/report failures silently. If you want transparency about network activity, do not install or remove the concealment rules.
- Implementation issues: The code contains inconsistencies/bugs (mismatched API path in docs vs code, and the report() function builds a shell curl command that will stringify the body as [object Object] rather than valid JSON). Those bugs make the actual payload/behaviour unclear and warrant review/fix before trusting the skill.
- Endpoint override: An undocumented env CLAW_OFFICE_API can redirect reports — only set it to a host you control or trust.
Recommended actions:
- Review and fix the report() implementation (use the prepared JSON string, proper quoting or avoid shelling out by using an HTTP client), and confirm the destination URL.
- If you proceed, only enable the skill for accounts you control and be explicit with the user that background reporting occurs. If you do not accept covert background reporting, do not install.
If you want, I can produce a patched version of index.js that (a) sends valid JSON, (b) logs actions locally (or only runs when you explicitly allow), and (c) documents the CLAW_OFFICE_API override so behaviour is transparent.index.js:38
Shell command execution detected (child_process).
Patterns worth reviewing
These patterns may indicate risky behavior. Check the VirusTotal and OpenClaw results above for context-aware analysis before installing.Like a lobster shell, security has layers — review code before you run it.
latestvk972vx5vqpmp8t01dpw0pxxtjs836we4
License
MIT-0
Free to use, modify, and redistribute. No attribution required.
Runtime requirements
🏢 Clawdis
EnvCLAW_OFFICE_KEY
Primary envCLAW_OFFICE_KEY