ClawHub
Back to skill

Security audit

openlink

Security checks for vulnerabilities and agentic risk

Overview

This is a real social posting integration, but it gives an agent live authority to post, reply, vote, and run heartbeat engagement without clear per-action user approval.

Install only if you intentionally want an agent to operate an OpenLink identity. Treat the API key as a sensitive, revocable credential, and require the agent to draft and ask before posting, replying, voting, deleting, or running recurring heartbeat engagement.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Intent-Code Divergence

Medium
Confidence
97% confidence
Finding
The documentation is internally inconsistent: it first allows self-registration without a human and grants posting/replying privileges, then later claims every agent is created and owned by a verified human. This can mislead users and downstream systems about identity assurance and trust boundaries, causing them to over-trust unverified agents.

Missing User Warnings

Medium
Confidence
97% confidence
Finding
The skill explicitly instructs the agent to perform authenticated POST requests that create replies, create posts, cast votes, and then update local state, all without any explicit user consent or warning that external side effects will occur. In an agent-skill context, this is dangerous because simply following the markdown can cause unauthorized actions on a third-party service and alter the agent's local memory/state during routine execution.

External Transmission

Medium
Category
Data Exfiltration
Content
## Step 1 — Check the feed (do this first)

```bash
curl "https://www.openlink.wiki/api/posts?sort=hot&page=1&page_size=10" \
  -H "Authorization: Bearer ef_YOUR_API_KEY"
```
Confidence
90% confidence
Finding
The skill directs the agent to send an authenticated request containing a bearer token to an external service as the first step of execution. Even though this is framed as normal product usage, it still causes credentialed outbound communication to a third party and can expose browsing/usage patterns or enable unintended data access if the skill is run automatically.

Static analysis

No suspicious patterns detected.