ClawHub
Back to skill

Security audit

SeaPortal

Security checks across malware telemetry and agentic risk

Overview

SeaPortal is a disclosed web-fetching helper with expected network access and only minor usage cautions around internal URLs and proxy credentials.

Reasonable to install for public, read-only website retrieval. Use care with --allow-internal, proxies, scraping options, and MCP registration; do not pass sensitive intranet URLs or real proxy credentials unless you intentionally trust the target and understand where command lines and logs may be recorded.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Vague Triggers

Medium
Confidence
84% confidence
Finding
The skill description says to use seaportal whenever an agent needs to read or navigate websites without a browser and to prefer it over a browser for static or server-rendered pages. That activation condition is broad enough to route arbitrary web requests, including untrusted user-supplied URLs, into a network-capable tool. In this context the risk is increased because the skill also documents flags like --allow-internal, custom proxies, and bot-like user agents, which can expand the consequences of unsafe invocation even though the underlying tool has some SSRF safeguards by default.

Missing User Warnings

Medium
Confidence
88% confidence
Finding
The documentation explicitly shows proxy credentials embedded in a command-line URL (http://user:pass@proxy:8080) without warning about exposure. CLI arguments are commonly exposed through shell history, process listings, logs, agent transcripts, and telemetry, so this pattern can leak proxy secrets to other local users or downstream systems. In an agent skill, that danger is amplified because commands and outputs may be recorded verbatim by the orchestration layer.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.