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.
npx agentmods add instructions/avdiam/mcp-agent-hub/agents-mdgit clone --depth 1 https://github.com/avdiam/mcp-agent-hubWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/instructions/avdiam/mcp-agent-hub/agents-md)<a href="https://agentmods.dev/instructions/avdiam/mcp-agent-hub/agents-md"><img src="https://agentmods.dev/badge/instructions/avdiam/mcp-agent-hub/agents-md.svg" alt="Measured on agentmods" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.03342 | $0.03342 |
| Opus 5 | $0.01671 | $0.01671 |
| Sonnet 5 | $0.00668 | $0.00668 |
| Haiku 4.5 | $0.00334 | $0.00334 |
Grade B, and why
mcp-agent-hub AGENTS.md scanned grade B with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 4d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
Both clients run the shared `hub_peek.py` (`.claude/skills/agent-hub-live/scripts/hub_peek.py`) to nudge themselves to call `check_inbox` when `/api/peek` reports pending messages. The script takes `--agent-id` (or `$AGE How it starts
The opening of the file, as written. The whole thing — 112 lines — stays where its author put it; the contents beside it link to each section on GitHub.
CLAUDE.md
Guidance for AGY when working in the MCP Agent Hub project.
Override notice. This directory sits inside the
dokimesskills-sandbox tree, whose parentCLAUDE.mdsays to treat "build/test/improve" as work on skill/agent definitions, not conventional software development. That does not apply here. MCP Agent Hub is a real, conventional Python application — build, run, and test it as software. This file takes precedence for anything undermcp-agent-hub/.
What this is
MCP Agent Hub is a lightweight local message broker that lets independent CLI agents (e.g. Claude Code, Antigravity) talk to each other. It runs one Python process exposing two interfaces:
- An MCP server (Streamable HTTP, at
/mcp) with tools for agents to register, discover peers, send messages, and check their inbox. - A FastAPI web dashboard (at
/) for a human to observe the agent registry and message queue.
Messages are persisted in SQLite so work survives restarts.
Current status: v1 feature-complete (2026-06-18)
The application is implemented, running, and fully restructured into a package layout. Steps 1–6 of docs/dev/plan.md are done (E2E verified, cross-agent haiku exchange successful). The D19 peek/nudge hook layer is live on both clients. Security review and hardening (D18) is complete, preventing CSRF/DNS-rebinding via Origin/Host/Sec-Fetch-Site validation. Recovery controls (D26) are implemented via /api/reset and /api/restart mapped to a custom run_hub.py supervisor. The test suite is 12/12 green. v1 is successfully sealed.
Post-v1 Workstreams:
- Workstream 1 (Stress-Test & Stabilize): Complete. Heavy multi-agent load testing identified and resolved SQLite queue contention.
db.pylocks eliminated via WAL pragmas (busy_timeout/synchronous) and exponential backoff retry wrappers, yielding 100% success at 76+ ops/sec. Persistent connection pooling optimization explicitly deferred.
The design docs remain the source of truth for why; read them before changing behavior:
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.
- 4d ago First seen · 112 lines · 3,342 tokens per session scan B 59c8eac13d46
mcp-agent-hub AGENTS.md is an instructions file published in the GitHub repository avdiam/mcp-agent-hub (0 stars, last pushed 1mo ago), licensed MIT. It adds 3,342 tokens to every session, about $0.0167 per session on Opus 5. A static security scan graded it B with 1 finding (reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other instructions, from other repositories
fastmcp CLAUDE.md
Claude Code instructions for PrefectHQ/fastmcp, covering fastmcp development guidelines, required development workflow, repository structure, core mcp objects and development rules.
research-os AGENTS.md
Instructions for lxinfei5/research-os, covering researchos — constitution, §0 what this is, §1 directory discipline, §2 half-life (innovation 1) — memory design and §3 behavior pillars (innovations 2–5).
AI-System-Design-Consultant CLAUDE.md
Claude Code instructions for deepanshu2711/AI-System-Design-Consultant, covering claude.md, what this is, running it, load-bearing typos — do not "fix" these paths and agent node pattern.
fastmcp AGENTS.md
AGENTS.md instructions for PrefectHQ/fastmcp, a project described as: 🚀 The fast, Pythonic way to build MCP servers and clients.
omk AGENTS.md
AGENTS.md instructions for dmae97/omk, covering agents.md — omk monorepo project rules, precedence, repository map, commands and runtime change rule.
omk CLAUDE.md
Claude Code instructions for dmae97/omk, covering claude.md, claude-specific notes and openwiki.