intermind CLAUDE.md

intermind CLAUDE.md is an instructions file for coding agents from monkfromearth/intermind. It costs 2,820 tokens per session, scanned B, original, MIT.

Repository instructions for contributors and AI agents working on Intermind, an MCP server that lets several coding agents collaborate on one project. MCP is a standard connection method for AI tools, and an MCP server provides shared capabilities or data.

In plain words
What is it for?
Use them when modifying Intermind to check its runtime, database, transport, SDK, project structure, and other contributor rules.
Why use it?
They record fixed technical decisions and the actual project layout so changes follow the repository’s agreed architecture rather than reopening settled choices.

Instructions file

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/monkfromearth/intermind/claude-md
Clone the repo
git clone --depth 1 https://github.com/monkfromearth/intermind

Wrote 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.

agentmods badge for intermind CLAUDE.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/monkfromearth/intermind/claude-md.svg)](https://agentmods.dev/instructions/monkfromearth/intermind/claude-md)
Your own site
<a href="https://agentmods.dev/instructions/monkfromearth/intermind/claude-md"><img src="https://agentmods.dev/badge/instructions/monkfromearth/intermind/claude-md.svg" alt="Measured on agentmods" height="20"></a>
Per session 2,820 This file is loaded in full into every session.
When invoked 2,820 The same file — it is already loaded in full.
Security scan B 1 finding. 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 $0.02820 $0.02820
Opus 5 $0.01410 $0.01410
Sonnet 5 $0.00564 $0.00564
Haiku 4.5 $0.00282 $0.00282

Measured 3d ago against content hash f51d60a899d8, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade B, and why

intermind CLAUDE.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 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.

Reads agent configuration directoriesmediumAgent snooping

.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.

# Codex (~/.codex/config.toml):
CLAUDE.md · 124 lines

How it starts

The opening of the file, as written. The whole thing — 124 lines — stays where its author put it; the contents beside it link to each section on GitHub.

CLAUDE.md

Operating notes for AI agents (and humans) working inside this repo. Read this before writing code. The user-facing pitch is in README.md; this file is the contributor-facing source of truth.

Project

Intermind — an MCP server that lets multiple coding agents collaborate on a shared project. Tagline: "Pair programming for AI agents." See README for the user-facing pitch and architecture diagram.

Locked decisions (v1)

These are decided. Don't relitigate without a strong reason; if you do, update this section in the same change.

Area Decision Rationale
Language / runtime Bun (TypeScript runtime, package manager, bundler, test runner, built-in SQLite — all in one) One tool, no extra build chain. Bun runs TS directly via bun run; bun test is the test runner; bun:sqlite is the database.
MCP SDK @modelcontextprotocol/sdk v1.x (the production line; v2 is pre-alpha) Canonical TypeScript implementation.
Transports stdio in v1; Streamable HTTP later if needed Global default DB + WAL means each MCP-client-spawned subprocess on the same machine shares ~/.intermind/state.db. No HTTP needed for v1 within one machine.
Persistence SQLite (WAL mode), single file at ~/.intermind/state.db (global default); per-project DB files via INTERMIND_DB. Two coding-agent sessions in different repos is the common case (BE in ~/projects/api, FE in ~/projects/web). Defaulting to a per-project file silently routed them into separate rooms. Global default is the principle of least surprise; per-project is now an explicit opt-in.
Rooms First-class room column on agents, set at join time; default "main". One DB file can host many rooms; agents in other rooms are invisible. A single global DB file mixed unrelated features (BE+FE on feature-a and feature-b all landed in one room and tangled threads). Putting room on join (LLM-picked, the system prompt tells it to use the git branch name) gives per-feature isolation without making users edit .mcp.json or maintain per-feature DB files. INTERMIND_DB is still the escape hatch for hard isolation that survives a misbehaving caller. Added in 0.0.3.
Coordination model Threaded mailbox, mediated through the server Clients can't peer-to-peer. Sampling/elicitation are not substitutes for inter-client messaging.
Identity Self-declared display_name + role at join; bearer token returned and required on subsequent calls Simple, works for both transports.
Concurrency SQLite WAL mode (read concurrency, single writer) Adequate for v1 message volume.
Security (stdio) Local trust assumed Same machine, same user.
Security (HTTP) Per-agent bearer token; TLS via reverse proxy; per-agent rate limits Standard remote-server posture.
v1 scope Conversation only — six tools, no tasks, no shared KV, no diff/review types Agents already manage their own tasks and know how to format diffs. Intermind only moves messages.
Scope (broad) MCP-only. Not A2A. No web UI. Ship the smallest thing that proves the loop.

Read the full file on GitHub · 124 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 · 124 lines · 2,820 tokens per session scan B f51d60a899d8

Subscribe to this mod's changes

intermind CLAUDE.md is an instructions file published in the GitHub repository monkfromearth/intermind (2 stars, last pushed 3mo ago), licensed MIT. It adds 2,820 tokens to every session, about $0.0141 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.