LangChain's Managed Deep Agents keeps credentials out of the sandbox via proxy-injected tokens
LangChain · x · 2026-10-17
LangChain details the credential-safety design of Managed Deep Agents:
- Problem: agents with sandboxes run packages you didn't write; one malicious package exposes every credential inside the sandbox.
- Solution: a managed proxy attaches your token outside the sandbox, only on calls to services you've explicitly allowed — the agent never sees it.
- Platform: you define an agent as a directory of files (instructions, tools, skills, middleware, MCP connectors); LangSmith Agent Server runs the harness and runtime, providing API, threads, streaming and an MCP endpoint with no infra to operate.
More from coding & agent
- AutoSep cuts agent reward hacking from 69.9% to 12.1% via weak-to-strong verifier hardening — Stanford · 2026-10-17
- Free open-source course teaches harness engineering with 14 lectures and 8 hands-on agent projects — techNmak · 2026-10-17
- Anthropic engineer: Claude Slides is now how I make slide decks — bcherny · 2026-10-17
- Dev built soft-body physics engine with Claude's help, claims it tops BeamNG — grandprixlegends · 2026-10-17
- tldraw ships Codex plugin for pair-programming on a live multiplayer canvas — max__drake · 2026-10-17
- Open Co-Scientist: open-source recreation of Google's AI Co-Scientist with Elo-ranked hypothesis evolution — Haunting-Stretch8069 · 2026-10-17