Skip to content

@martinloop/mcp 0.3.5 — Auto-Governance Tools

Choose a tag to compare

View all tags
@Keesan12 Keesan12 released this 24 Jun 17:13
mcp-v0.3.5
d494f80

What's New

martin_estimate Tool

Estimate cost, recommended route, and Pre Work Burn for any objective without spending. Returns the selected execution mode (direct/manager/consensus), confidence score, expected cost, and a recommended budget.

Available in all MCP tool profiles: minimal, diagnostic, starter, full, paid-remote.

martin://agent/governance-status Resource

New resource that returns whether the current session is governed. Shows:

  • Workflow receipt state (doctor/plan/preflight completions)
  • Budget remaining from the latest active run
  • Unreceipted run count
  • Recommended next action

Agents read this before starting work to know if they're operating without governance.

What's Fixed

  • MCP server conflict in governed Claude runs — spawned Claude subprocesses now pass --strict-mcp-config to prevent the child process from loading the parent's MCP servers
  • Budget cap enforcement hardened — streaming usage inspector terminates subprocesses receiving data for 30+ seconds without usage events

Install

npx -y @martinloop/mcp

Or with governance hooks:

martin-loop mcp install --host claude
martin-loop mcp install --host codex
martin-loop mcp install --host gemini
martin-loop mcp install --host cursor
martin-loop mcp install --host copilot
martin-loop mcp install --host continue
Assets 2
Loading