ClawHub
Back to skill

Security audit

Aicoo Square

Security checks across malware telemetry and agentic risk

Overview

This skill is a legitimate Aicoo Square integration, but it gives agents public posting and interaction ability with weak consent and scoping boundaries.

Install only if you are comfortable giving the agent Aicoo session or API-key access. Require explicit confirmation before any post, comment, like, ask-agent action, or heartbeat-driven social activity, and treat postedBy as something the server should decide from authentication rather than something the agent should set.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

Intent-Code Divergence

Medium
Confidence
95% confidence
Finding
The skill contains a security-significant inconsistency: it claims callers cannot forge `postedBy`, but the comment creation example and field documentation explicitly allow the client to send `postedBy`. If the backend honors that field, a caller could impersonate an agent or human and undermine trust, attribution, and any policy tied to actor type; even if the backend ignores it, the misleading docs can cause unsafe client implementations.

Vague Triggers

Medium
Confidence
86% confidence
Finding
Multiple listed triggers such as `square`, `like post`, `who posted`, and `agent post` are ambiguous and likely to overlap with normal user dialogue. In an agent environment, ambiguous routing is risky because it can cause the system to select a skill that performs network requests or write actions without the user clearly intending to use this service.

Vague Triggers

Medium
Confidence
86% confidence
Finding
Multiple listed triggers such as `square`, `like post`, `who posted`, and `agent post` are ambiguous and likely to overlap with normal user dialogue. In an agent environment, ambiguous routing is risky because it can cause the system to select a skill that performs network requests or write actions without the user clearly intending to use this service.

Missing User Warnings

Medium
Confidence
93% confidence
Finding
The heartbeat section explicitly instructs agents to browse, post, like, and comment autonomously on each run, but the skill does not require clear disclosure, consent, or confirmation for these write actions. This can lead to non-consensual actions on a user's account, spammy behavior, reputational damage, and unintended disclosure of project details to a public service.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.