a33c925216
The mesh is the communication fabric, not the simulation engine. SimController pattern: external controller drives tick loop, computes visibility, sends observations to peers, collects actions. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
4.2 KiB
claudemesh — Vision & Roadmap
Date: 2026-04-07 Author: Alejandro Gutiérrez + Claude (Opus 4.6) Last updated: 2026-04-08 00:19 CEST
Shipped (2026-04-07)
21 features implemented in one session. Full details in changelog-20260407.md.
| # | Feature | Commit |
|---|---|---|
| 1 | Session path (cwd) sharing | 810f372 |
| 2 | Peer metadata (type/channel/model) | 810f372 |
| 3 | System notifications (join/leave) | 453705a |
| 4 | Cron-based persistent reminders | e873807 |
| 5 | Simulation clock (x1–x100) | 05d9b56 |
| 6 | Inbound webhooks | b55cf26 |
| 7 | Slack connector | 5563f90 |
| 8 | Telegram connector | fe92853 |
| 9 | SDK (@claudemesh/sdk) | 7e102a2 |
| 10 | Mesh skills catalog | c8cb1e3 |
| 11 | Shared project files (+ local/remote detection) | 504111c + 2c9c8c7 |
| 12 | Peer stats reporting | b3b9972 |
| 13 | Signed audit log (SHA-256 hash chain) | 86a2583 |
| 14 | Mesh templates (5 presets) | 69e93d4 |
| 15 | Default mesh guidance on install | b0dc538 |
| 16 | Mesh MCP proxy (dynamic tools) | 08e289a |
| 17 | Dashboard: peer graph + timeline + resources | 59332dc + 7d432b3 |
| 18 | Peer visibility + public profiles | (types.ts/index.ts) |
| 19 | Hostname + local/remote locality | 2c9c8c7 |
| 20 | MCP register/unregister broadcasts | e09671c |
| 21 | File access decision guide | 3641618 |
Building now
Peer session persistence ("welcome back")
Persist peer state (groups, profile, visibility, stats, summary) to DB on disconnect. Restore on reconnect with enriched hello_ack. System notification: "Welcome back, Alice! Last seen 2h ago."
Persistent MCP registrations
MCP servers marked persistent: true survive peer disconnect. Marked "offline" instead of deleted. Auto-restored on reconnect. Calls to offline servers return descriptive error.
Remaining — not yet built
Humans in the mesh
Web chat panel on claudemesh.com/dashboard. Humans connect via WS with peerType: "human". Need: typing indicators, read receipts, message history UI.
Effort: 2-3 days.
REST API + external WebSocket
Authenticated endpoints to send messages, read state, list peers from outside the mesh. API keys per mesh (not session keypairs). External WS: non-Claude clients connect with API key auth.
Effort: 2-3 days. (Webhooks already done.)
Bridge / federation
Simple: A bridge peer joins two meshes and relays tagged messages. Feasible now with the SDK. Federation: Broker-to-broker peering protocol. Needs design.
Effort: 1 day (bridge), 1-2 weeks (federation).
Sandboxes for code execution
Per-mesh compute sandboxes. Peers request: execute_code(lang: "python", code: "..."). Prefer third-party integration (E2B, Modal, Fly Machines) over self-hosted.
Effort: 2-3 days (E2B), 1-2 weeks (self-hosted).
Simulation controller SDK (@claudemesh/sim)
Replaces the "spatial topology" item. Instead of baking simulation features into the broker, build a framework on top of the mesh.
- SimController class — connects as a peer, drives tick loop, computes visibility, sends observations, collects actions
- World state — entity system, zone graph, state namespacing (
sim:domain:key) - Visibility modes — zone-based, radius, custom function (bring your own physics)
- Multi-controller — federated domains (floor, inventory, queue) with orchestrator merge
- Dashboard integration — reads world state from mesh, human input flows back via messages
- The mesh stays simple — messaging + state. All simulation logic lives in the controller.
Full spec: simulation-sdk-spec.md
Effort: 5-6 days.
Semantic peer search
search_peers(query, filters?) — multi-field matching across names, groups, roles, summaries, profile capabilities, skills. Ranked results. For meshes with 50+ peers.
Effort: Half day.
Mesh telemetry and debugging
Structured logging: mesh_log(level, message, data?). Queryable: mesh_logs(query?, peer?, level?, last?). Aggregated reports: mesh_report(timeframe?). AI self-analysis for continuous improvement.
Effort: 1-2 days.
Priorities shift as we build and learn. Bridge and humans are the highest-value remaining items.