An agent tool is a permission boundary
Look past the tool description. What can the agent actually do, with whose identity, and under which constraints?
Agentic / MCP Security Series.
When AI can take action, the boundaries matter more. Study agents, tool permissions, and the Model Context Protocol.
Articles in this series
Isolating Tool Output in Agents
Keep information from tools separate from permission to act, with clear examples of provenance, narrow tools, and independent authorization.
What this series explores
- 01Tool authorization
- 02MCP trust boundaries
- 03Agent isolation
- 04Human approval design