Agent Security: Policy-Driven Gateway for Tool Discovery
Strange_Profit_8129 · reddit · 2026-08-20
Addressing the security risk where most MCP setups expose tools/list widely (allowing agents to see schemas for tools they can't invoke), the author implemented a gateway solution.
Core Mechanism:
- Ties discovery to the same policy engine used for authorization.
- Different callers receive different tool catalogs when hitting the same endpoint.
- Ensures unauthorized tools (e.g., databasedelete) are hidden at the discovery stage.
The gateway also proxies prompts and resources across multiple upstreams and is available as a Spring Boot Starter (MIT licensed).
More from coding & agent
- Tempo explores Human Authorization in Agentic Workflows — soumitrashukla9 · 2026-08-20
- Open source AI assistant plugin for WordPress launches with Agent UI — Scobleizer · 2026-08-20
- Developer switches from Claude Code to FactoryAI — matanSF · 2026-08-20
- Building an infrastructure layer for AI agents in Meta Ads automation — Civil-Historian9568 · 2026-08-20
- DSRs docs integrate mixedbreadai's Toast-1 for AI support — krypticmouse · 2026-08-20
- FrankenRedis: A memory-safe Rust reimplementation of Redis, built with AI agents — doodlestein · 2026-08-20