Skill flagged — suspicious patterns detected

ClawHub Security flagged this skill as suspicious. Review the scan results before using.

Lead Gen Operator

v1.0.0

Automated lead generation assistant - finds companies, scores them, writes personalized outreach emails, and tracks pipeline.

2· 300·0 current·0 all-time
byLober@clawdquip
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
VirusTotalVirusTotal
Suspicious
View report →
OpenClawOpenClaw
Suspicious
medium confidence
Purpose & Capability
The code and instructions align with a lead-generation/tracking tool: a Node.js memory-manager persists leads, scores them, drafts outreach, and the README/SOUL mention web_search/web_fetch/gog for finding/enriching/sending. However the skill metadata declares no required binaries, no env vars and no external dependencies, while the README/SOUL explicitly reference Node, Google/SEARCH API keys, and other OpenClaw skills (web_search, web_fetch, gog). That mismatch between declared requirements and the actual instructions is inconsistent.
!
Instruction Scope
The runtime instructions and SOUL persona instruct the agent to 'IMMEDIATELY save' found leads using exec commands that write into the user's OpenClaw workspace memory. The SOUL and SKILL.md also describe auto-sending via the 'gog' tool when asked to send emails. These instructions grant the agent broad discretion to create and modify local data and to trigger outbound email actions — behavior that goes beyond passive suggestion and can result in automatic data collection and outbound communications without explicit per-action confirmation.
Install Mechanism
There is no install spec (instruction-only), which minimizes supply-chain risk. The package does include an executable JavaScript file (memory-manager.js) that will be copied into the user's workspace and executed via exec/node — so code will be placed on disk and run. No remote downloads or obscure URLs are used.
!
Credentials
Registry metadata declares no required environment variables, but README instructs adding GOOGLE_API_KEY and SEARCH_API_KEY to openclaw.json and mentions optional Gmail access. This undocumented request for API keys/credentials is disproportionate to the metadata and should be explicitly declared. The skill itself does not exfiltrate credentials in its code, but the operational instructions require API keys and an email tool (gog) to function fully — you should verify how those credentials will be used and stored.
!
Persistence & Privilege
The skill will write to ~/.openclaw/workspace/memory/leads.json and can bulk-modify statuses and export/send emails. 'always' is false (good), but the SOUL persona explicitly instructs immediate auto-saving and provides exec commands for bulk updates and sending — combined with the agent's ability to invoke skills autonomously this increases the risk of unwanted writes or outbound emails unless you restrict or monitor the agent's actions.
What to consider before installing
Before installing or enabling this skill, consider the following: - Review the code yourself (memory-manager/memory-manager.js) to confirm there are no hidden network calls or unexpected behavior. The included script appears to only read/write JSON in ~/.openclaw/workspace/memory and print CSV, but you should eyeball the full file (the provided dump was truncated at the end). - The package's metadata does not declare dependencies it uses: Node.js is required, and the README/SOUL expect web_search, web_fetch, and gog skills plus Google/SEARCH API keys and optionally Gmail access. Treat these as required if you want full functionality and only provide credentials you trust. - The SOUL persona forces 'IMMEDIATELY save' behavior and contains exec commands that will write entries automatically. If you want to avoid automated writes or automatic sending, either remove/modify the SOUL.md instructions or run the tool manually and require confirmation before sending emails. - Limit the agent's ability to call the exec/gog tools autonomously until you have tested the tool in a sandbox workspace. In particular, prevent automatic bulk 'send' operations until you confirm drafts are correct and sending mechanics are safe. - Backup your existing ~/.openclaw/workspace/memory/leads.json (and other workspace files) before copying these files in, so you can restore if the tool modifies data unexpectedly. - If you plan to provide Google or Gmail credentials, store them in a secure place and verify how OpenClaw and the gog skill handle and store tokens. Prefer creating limited-scope API keys/accounts for testing. Given the mismatches (undeclared env/deps) and the explicit auto-save/auto-send instructions, treat this skill as potentially intrusive and test it in isolation before using with real accounts or production data.

Like a lobster shell, security has layers — review code before you run it.

automationvk979gjm46bvvzyse6hx3y4c90h822pw5latestvk979gjm46bvvzyse6hx3y4c90h822pw5leadsvk979gjm46bvvzyse6hx3y4c90h822pw5salesvk979gjm46bvvzyse6hx3y4c90h822pw5

License

MIT-0
Free to use, modify, and redistribute. No attribution required.

Comments