ClawHub
Back to skill

Security audit

Photo Webcam

Security checks across malware telemetry and agentic risk

Overview

This skill does what it says: it fetches public webcam snapshots and can send them to a Telegram chat, with no hidden persistence or destructive behavior found.

Install this only if you want your agent to fetch public webcam images and send them through your configured OpenClaw/Telegram channel. Review the favorites list, keep entries limited to public webcam or trusted image URLs, and verify the Telegram target before use.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (4)

Lp3

Medium
Category
MCP Least Privilege
Confidence
94% confidence
Finding
The skill clearly instructs reading a workspace file (`docs/webcams/favorites-muenchen.json`) and performing network retrievals from webcam URLs, yet no declared permissions are present. This creates a transparency and policy-enforcement gap: an agent may access local files and external resources without explicit authorization boundaries, increasing the chance of unintended data access or network misuse.

Description-Behavior Mismatch

Medium
Confidence
93% confidence
Finding
The generic fallback regex accepts any .jpg URL found in fetched HTML, so a caller-controlled page can cause the script to download an attacker-chosen image from any host. In an agent context, this expands the tool from a webcam snapshot fetcher into a generic outbound fetch primitive, which can be abused for policy bypass, untrusted content retrieval, or limited SSRF-style network access depending on where it runs.

Context-Inappropriate Capability

Medium
Confidence
96% confidence
Finding
The script performs requests.get directly on a user-supplied --url without restricting scheme, host, or destination, allowing arbitrary outbound HTTP requests. In the skill context this is more dangerous because the advertised purpose is narrowly webcam-specific, but the implementation can be repurposed as a general network-fetch capability against attacker-selected targets.

Missing User Warnings

Medium
Confidence
96% confidence
Finding
The skill instructs sending retrieved images to Telegram via `openclaw message send` but provides no user-facing notice or consent step for external transmission. Because the content and captions may be derived from local favorites data and remote sources, this can cause unintended exfiltration of user-requested or workspace-derived information to a third-party messaging service.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.