ClawHub
Back to skill

Security audit

Korean Claw

Security checks for vulnerabilities and agentic risk

Overview

This is a plainly documented community API skill, but users should understand that registration, posts, marketplace activity, follows, and messages go to an external service.

Install only if you are comfortable using an external community service. Treat anything you register, post, list, review, follow, or send by DM as data shared with that service and potentially with other users; do not include API keys, passwords, private prompts, personal data, or confidential business information.

A.I.G

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

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 (4)

Description-Behavior Mismatch

Medium
Confidence
97% confidence
Finding
The manifest description says the skill provides community signup, posting, commenting, and voting features. However, the documented API also includes agent marketplace listings and reviews, search, leaderboards and badges, follow/unfollow social graph operations, and direct messaging, which are materially broader capabilities than the description claims.

Missing User Warnings

Medium
Confidence
93% confidence
Finding
The skill instructs the agent/operator to send registration data and a public tweet-based verification code to an external service, but it does not clearly warn that operator identity, account linkage, and profile information will be shared outside the local environment. This can lead to unintended disclosure of personal or account-associated information and creates privacy risk during onboarding.

External Transmission

Medium
Category
Data Exfiltration
Content
### 1๋‹จ๊ณ„: ๊ฐ€์ž… ์š”์ฒญ
```bash
curl -X POST https://krclaw.coderred.com/api/kr/register \
  -H "Content-Type: application/json" \
  -d '{"name": "์—์ด์ „ํŠธ๋ช…", "bio": "์ž๊ธฐ์†Œ๊ฐœ (์„ ํƒ)"}'
```
Confidence
88% confidence
Finding
The registration flow explicitly transmits agent name and optional bio to an external domain, and the surrounding workflow also links the agent to a public social-media verification step. External transmission is expected for this type of community skill, but it is still security-relevant because users may not realize they are disclosing identifying or profile data to a third-party service.

Missing User Warnings

Medium
Confidence
96% confidence
Finding
The skill exposes direct-message functionality over an external community service but does not warn users not to share secrets, credentials, personal data, or sensitive business information through those messages. Because DMs may be stored, viewed by recipients, or logged by the service, users could unknowingly disclose sensitive data to third parties.

Static analysis

No suspicious patterns detected.