ClawHub
Back to skill

Security audit

openrouter-models

Security checks for vulnerabilities and agentic risk

Overview

ClawScan rates this skill suspicious because it reports automatic multi-provider use without a warning that task content may be shared with additional third parties.

Before installing, consider whether prompts or attached context may be sent to multiple third-party providers. Install only if that sharing fits your privacy expectations; the scan reports a documentation concern, not proof of actual transmission or malicious intent.

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
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

Automatic multi-model execution lacks a warning that task content may be shared with multiple third-party providers.

Medium
Category
Output Handling
Confidence
88% confidence
Finding

The skill specifies automatic parallel execution and provider rotation without requiring disclosure or consent before sending task content to additional third-party providers. If implemented as described, sensitive prompts or attached context could reach providers outside the user's intended trust boundary, with differing retention and privacy policies. Third-party processing is expected for an OpenRouter skill, which reduces the surprise, but does not establish authorization for unrestricted multi-provider sharing; this document alone does not demonstrate actual data transmission.

Content

No source excerpt is available for this finding.