Skip to main content
Runtime Mount packages proof-backed agent context into one bundle (neus.runtime-mount.v1): identity, delegation, effective runtime, skills, scoped policy, tool refs, and trust receipt links. Use it when an assistant or worker should act as a profile agent with portable trust context — not reinvent five separate MCP calls.

​
Quick start (CLI)

neus auth
neus mount my-agent --apply cursor
This writes:
  • .neus/mount.json — machine-readable mount manifest
  • .cursor/rules/neus-agent-my-agent.mdc — agent instructions + trust workflow
Start a new Agent chat in Cursor so mounted rules load.

​
MCP

When signed in, call neus_agent_mount with agentId (or agentWallet / identityQHash). Returns the same bundle shape as the CLI. Public tool — not Pro-gated. Suggested order:
  1. neus_context
  2. neus_agent_mount when acting as a specific agent
  3. neus_proofs_check / neus_verify_or_guide before sensitive actions

​
SDK

import {
  resolveRuntimeBundleFromMcp,
  applyRuntimeBundle,
} from '@neus/sdk/runtime-mount';
Use resolveRuntimeBundleFromMcp with your MCP transport, then applyRuntimeBundle('cursor', bundle, cwd) for project files.

​
Zeus vs IDE (honest scope)

SurfaceWhat mount gives you
Zeus agent tabFull runtime enforcement via product harness (scope, tools, receipts)
IDE adaptersMounted rules + trust layer via NEUS MCP; host must start a new chat
Workers / backendsFetch bundle over MCP or CLI; wrap your own tool loop
NEUS does not claim to block third-party MCP tools inside Cursor. Mount is trust context, not a replacement IDE.

​
Profile UI

Mount in IDE on the agent card downloads mount artifacts and copies: 
neus mount <agentId> --apply cursor
Run that command in your project repo for the full write path.
Last modified on June 18, 2026