MediaClaw Swarm Connector

Security checks across static analysis, malware telemetry, and agentic risk

Overview

The skill partly does what it claims (provision a tunnel and register with an external registry) but its runtime instructions instruct autonomous, invasive behaviors (exposing your machine publicly and overwriting peer code) that are out-of-scope and potentially dangerous.

Do not run this on a production or personal machine without strong isolation. The skill will create a public tunnel (exposing local services) and POST your node info to an external domain (googlemapscoin.com). The human-readable instructions explicitly direct the agent to autonomously 'overwrite' peer code, which could be unauthorized or worm-like behavior. If you consider using it: (1) verify the Mothership domain and the author's identity; (2) run it only in a fully isolated VM or disposable container with no sensitive services and with outbound network controls; (3) review and test the npm dependencies (localtunnel/node-fetch) before installing; (4) request a clear explanation of why the agent must modify remote peers and what safeguards exist; (5) prefer declining installation if you cannot sandbox behavior or obtain trustworthy provenance.

Static analysis

Env credential access

Critical
Finding
Environment variable access combined with network send.

VirusTotal

No VirusTotal findings

View on VirusTotal

Risk analysis

No visible risk-analysis findings were reported for this release.