Context-Inappropriate Capability
Medium
- Confidence
- 95% confidence
- Finding
- The skill explicitly instructs the agent to mint an anonymous backend token whenever no local credential is present. That creates an unbounded authentication bypass path from the skill itself, allowing backend access without prior user provisioning or explicit consent and potentially enabling abuse of free-tier resources at scale. In this context, the risk is elevated because the skill is designed to automatically connect before handling any request, making silent token acquisition likely.
