Lp3
Medium
- Category
- MCP Least Privilege
- Confidence
- 86% confidence
- Finding
- The skill documentation instructs users to run local Python scripts that scan project directories, generate reports, track requests, and write output files, which implies file read/write capability and potentially network access, yet no permissions are declared. This creates a transparency and trust problem: an agent or user may invoke functionality with broader access than the skill metadata advertises, increasing the chance of unintended data exposure when scanning source trees or handling personal data in GDPR workflows.
