Skill flagged — suspicious patterns detected

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

Facebook Local Lead Sniper

v1.0.0

Automates joining and engaging in local Facebook groups to post recommendation requests, analyze replies, and generate lead outreach messages.

0· 73·0 current·0 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
Capability signals
CryptoCan make purchases
These labels describe what authority the skill may exercise. They are separate from suspicious or malicious moderation verdicts.
VirusTotalVirusTotal
Suspicious
View report →
OpenClawOpenClaw
Suspicious
high confidence
!
Purpose & Capability
The scripts and SKILL.md consistently implement Facebook group automation (join, like, comment, post, analyze, DM outreach) which matches the name/description. However the skill calls external binaries (curl, python3) and expects the web-access CDP proxy and a CLAUDE_SKILL_DIR layout, yet the registry metadata lists no required binaries or credentials — that mismatch is unexpected and should have been declared.
!
Instruction Scope
Runtime instructions and the scripts require control of your real Chrome session via a local CDP proxy (localhost:3456) and will open Facebook tabs, click UI elements, post in groups, comment, join groups, take screenshots, and generate/send outreach messages. While all these actions are within the skill's advertised purpose, they operate on your live account and can perform network-visible actions (posts/DMs) without additional explicit run-time confirmations — this is sensitive and could cause account suspension or unwanted outreach if used unintentionally.
Install Mechanism
No remote install steps or downloads are specified; the skill is distributed as scripts and templates in-repo. That limits remote-code install risk. (Note: SKILL.md is 'instruction-only' in registry but the package contains executable scripts — not an install network fetch.)
!
Credentials
The skill declares no required environment variables or credentials, yet at runtime it relies on: a web-access CDP proxy (localhost:3456), the CLAUDE_SKILL_DIR environment variable convention, and optionally CDP_PROXY_URL. It also invokes curl and python3 but those are not listed as required binaries. The absence of explicit declared dependencies (binaries and required local services) is an incoherence that can break runs or mask privilege assumptions.
Persistence & Privilege
always:false and no skill-wide persistent privileges are requested. The skill does operate against your existing browser session (so it has effective access to whatever your browser is logged into), but it does not claim to auto-enable itself or modify other skills. Autonomous invocation (agent can invoke) is default and not flagged by itself — combine that with the above concerns when deciding to allow autonomous runs.
What to consider before installing
Before installing or running this skill: (1) Understand it will control your real Chrome session (open tabs, click, post, comment, join groups, take screenshots) using the local CDP proxy — it can post or send DMs from whatever FB account is logged in. (2) Confirm you have and trust the 'web-access' CDP proxy running on localhost:3456, and that you have curl and python3 available; the skill does not declare those dependencies but calls them. (3) Consider using a throwaway or test Facebook account (not your primary) to evaluate behavior, since automated posting/commenting can trigger Facebook policy enforcement or cause reputational harm. (4) Review the DM templates and outreach flow — this skill is designed for lead generation and includes sales outreach text that will be sent to third parties. (5) Ask the author for provenance and an explicit dependency list (curl, python3, web-access version, required env vars like CLAUDE_SKILL_DIR or CDP_PROXY_URL) and for an option to require interactive confirmation before any posting or DMs. Providing those items or adding explicit runtime prompts would reduce the current concerns.

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

latestvk97d0bzr3pdcr1y0sqfmp5vdtd84a34p

License

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

Comments