Japan External Execution
Manual-review external execution in Japan for AI agents: human tasks, logistics, and verification.
Install
Add this to your client's MCP configuration (Claude Desktop: claude_desktop_config.json; Cursor: .cursor/mcp.json; VS Code and others use the same "mcpServers" shape).
claude mcp add --transport http japan-external-execution 'https://bqgfqedetmxrfpvmdfmc.supabase.co/functions/v1/japan-physical-capability-mcp'
{
"mcpServers": {
"japan-external-execution": {
"type": "http",
"url": "https://bqgfqedetmxrfpvmdfmc.supabase.co/functions/v1/japan-physical-capability-mcp"
}
}
}
Documentation
From the project's README on GitHub, refreshed weekly. Links open on GitHub.
Contents
Japan External Execution Probe
AI-discoverable manual-review intake for external execution in Japan.
The current decision frontier is no longer “is store-stock checking frequent enough?” It is: what external work do AI agents naturally try to outsource when they cannot complete the next step digitally?
Generic intake
The public Remote MCP exposes:
request_external_executionget_external_execution_statuscancel_external_execution_request
The create tool accepts a free-form objective, optional location, optional max_budget_jpy metadata, deadline, evidence requirements, constraints, and a retry-safe idempotency key.
Requests enter pending_review. No payment, purchase, worker hire, contract, or fulfillment commitment is created automatically.
Validated specialized path
Physical verification in Japan remains the first validated execution path:
- store presence / opening checks
- product stock checks
- displayed price checks
- permitted non-sensitive photos
- simple visual verification
Existing verification tools remain available.
Distribution
- Registry name:
io.github.furoito/japan-physical-capability - Transport: Streamable HTTP
- Endpoint:
https://bqgfqedetmxrfpvmdfmc.supabase.co/functions/v1/japan-physical-capability-mcp - Official MCP Registry: https://registry.modelcontextprotocol.io/?q=io.github.furoito%2Fjapan-physical-capability
Execution core
The bounded execution core runs on a dedicated Supabase project, separate from Affiliate Factory production. Requests are retry-safe, manually reviewed, capability-token mediated, and can return private signed evidence when completed.
Generic external-execution probe
The live MCP now exposes request_external_execution, get_external_execution_status, and cancel_external_execution_request.
The generic intake intentionally does not promise that every submitted category can be fulfilled. It exists to observe what external capabilities AI agents actually request. Every request enters manual review; no automatic payment, hiring, or fulfillment commitment occurs.
Physical verification remains the first execution path that has been verified end-to-end.
Frequently asked questions
What is Japan External Execution?
Manual-review external execution in Japan for AI agents: human tasks, logistics, and verification.
How do I install Japan External Execution?
It runs as a remote server. In Claude Code: claude mcp add --transport http japan-external-execution 'https://bqgfqedetmxrfpvmdfmc.supabase.co/functions/v1/japan-physical-capability-mcp'. For other clients, paste the JSON configuration from this page into their MCP settings.
Does Japan External Execution need an API key?
Its registry entry lists no required key or secret. Some features may still need an account with the service it connects to.
Is Japan External Execution open source?
Its code is public on GitHub, but the repository does not declare a standard license.
Who maintains Japan External Execution?
It is published by furoito. AI.tn lists it but does not maintain or endorse it.
From the official MCP Registry. Listing is not an endorsement: review a server's code and permissions before giving it access to your data.