Lp3
Medium
- Category
- MCP Least Privilege
- Confidence
- 97% confidence
- Finding
- The skill instructs the agent to use network access, accept secrets, write screenshots to disk, and read environment-variable overrides, but it does not declare permissions to make those capabilities explicit. That creates a trust and review gap: a caller may approve or invoke the skill without realizing it can transmit API keys to external services and persist potentially sensitive device screenshots locally.
