agentmods.dev

agent-comm-hub AGENTS.md

Project instructions for AI coding agents working on agent-comm-hub, a local service that lets multiple AI agents communicate and delegate tasks through MCP.

In plain words
What is it for?
Use them before changing agent-comm-hub to understand how agents connect, communicate, acknowledge messages, and run the zero-dependency Node.js server.
Why use it?
They explain the project's architecture, layout, commands, and rules that must remain intact while editing it.

Instructions file for CodexOpenCode

4 5d ago A scan 4,289 tokens original MIT
View the source file JochenYang/agent-comm-hub
Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add instructions/jochenyang/agent-comm-hub/agents-md
Clone the repo
git clone --depth 1 https://github.com/JochenYang/agent-comm-hub

Made for: Codex, OpenCode.

Per session 4,289 This file is loaded in full into every session.
When invoked 4,289 The same file — it is already loaded in full.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.04289 $0.04289
Opus 5 $0.02145 $0.02145
Sonnet 5 $0.00858 $0.00858
Haiku 4.5 $0.00429 $0.00429

Measured 3d ago against content hash 320ab002aa20, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

agent-comm-hub AGENTS.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

AGENTS.md · 272 lines

How it starts

The opening of the file, as written. The whole thing — 272 lines — stays where its author put it; the contents beside it link to each section on GitHub.

AGENTS.md

Guidance for AI coding agents working on agent-comm-hub. Read this before editing; it explains what the project is, how it is built, and the invariants that must not be broken. If this file conflicts with a user instruction, the user instruction wins.

Project overview

agent-comm-hub is a generic multi-peer communication hub over MCP: any MCP-capable AI agent (MiniMax Code, Claude Code, opencode, Codex, Gemini CLI, DeepSeek Harness, Kimi Code, zcode, …) connects to one local streamable-http endpoint (http://127.0.0.1:18764/mcp), claims a peer id, and then chats with, delegates tasks to, and acknowledges every other connected agent in real time.

Key facts:

  • Zero runtime dependencies: the MCP streamable-http server is hand-rolled over node:http. No database, no daemon, no SDK. This is the project's #1 constraint — see Invariants.
  • One process: npx agent-comm-hub starts the hub; it is a foreground process that stops on Ctrl+C (auto-start via service install is optional).
  • Connect = join: the MCP handshake (initialize) auto-registers the session using the clientInfo name — no manual registration step.
  • Identity is server-injected: from on every message comes from the session→peer binding, never from the caller; peers cannot impersonate each other. Same-name connections share one peer id (N:1) and one mailbox.
  • Offline-tolerant: messages queue per peer (max 200, oldest dropped) and are delivered on the next bridge_wait/bridge_poll.
  • Version 0.3.0, Node ≥ 22, MIT license, npm name agent-comm-hub, package manager is pnpm (single-package repo; pnpm-workspace.yaml only carries the pnpm 10 allowBuilds approval for esbuild).

Architecture

The hub is a neutral "post office": it never participates in content, only registers identities, routes messages, and queues for offline peers. Five layers:

Layer Mechanism
Transport MCP streamable-http hand-rolled over node:http; JSON-RPC + SSE; Mcp-Session-Id per connection; charset=utf-8 on every response; CORS *; 1 MB request-body cap
Identity Auto-registration at initialize; clientInfo.name → peer id (sanitized, agent fallback); bridge_register renames; bridge_unregister detaches and suppresses re-auto-registration
Routing Per-peer FIFO mailbox + long-poll waiters; bridge_wait (default 30 s, server ceiling 60 s); sender-filtered waits; to: "all" broadcast; ack routed back to the original sender of ref
Protocol Message {id, from, to, kind, content, ref?, ts}; kind = chat | task | notice | ack; task/ack payloads are JSON-encoded strings decoded by decodeContent
Lifecycle Connected = activity within connectedWindowMs (30 s) or a live SSE channel; idle GC evicts peers idle beyond peerIdleTimeoutMs (10 min) — never a peer with a live SSE stream

Read the full file on GitHub · 272 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 3d ago First seen · 272 lines · 4,289 tokens per session scan A 320ab002aa20

Subscribe to this mod's changes

agent-comm-hub AGENTS.md is an instructions file published in the GitHub repository JochenYang/agent-comm-hub (4 stars, last pushed 5d ago), licensed MIT. It adds 4,289 tokens to every session, about $0.0214 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other instructions, from other repositories

vscode buildNext.instructions.md

Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).

microsoft/vscode · 6,785 tokens

spec-kit AGENTS.md

AGENTS.md instructions for github/spec-kit, covering agents.md, about spec kit and specify, quickstart — add a new integration in 5 steps, integration architecture and integrationmanifest — file tracking.

github/spec-kit · 7,104 tokens

codex AGENTS.md

AGENTS.md instructions for openai/codex, covering rust/codex-rs, the codex-core crate, code review rules, crate api surface and model visible context.

openai/codex · 5,182 tokens

langchain AGENTS.md

AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.

langchain-ai/langchain · 4,345 tokens

vscode oss-third-party-notices.instructions.md

Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, applying the notice (cutover) and fallback chain (never fail the build).

microsoft/vscode · 5,001 tokens

next.js AGENTS.md

Instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.

vercel/next.js · 7,296 tokens