Local-first emotional-continuity MCP server: right shape or should this stay a CLI?

[pioneerjeff]

Pre-submission Checklist

• This post follows the MCP community guidelines

What would you like to share?

Hi, quick design question from an Emotion Engine v0.2.2 release.

I added a local stdio MCP server for runtime/state tools, plus registration examples for Codex CLI, Claude Code, Claude Desktop, and project-level .mcp.json.

The server is intentionally local-first: it reads/writes a project-local JSON state file and does not host user state remotely.

For MCP-capable local agents, does this feel like the right shape?

More specifically: would you expect an emotional-continuity layer to be exposed as an MCP server, or would a plain repo-local CLI/helper script be more natural?

Relevant Links

• Release: https://github.com/pioneerjeff-labs/emotion-engine/releases/tag/v0.2.2
• MCP docs: https://github.com/pioneerjeff-labs/emotion-engine/blob/main/docs/MCP.md
• Demo: https://pioneerjeff-labs.github.io/emotion-engine/demo/