Lp3
Medium
- Category
- MCP Least Privilege
- Confidence
- 90% confidence
- Finding
- The skill instructs the agent to read multiple local reference files and to invoke a local script (`check_reporting.py`), which are code-like capabilities, but no corresponding permissions are declared. This creates a trust and containment gap: a host may expose file access implicitly or users may not understand that the skill depends on local file reads and command execution to function, increasing the chance of unintended data access or unsafe deployment assumptions.
