What is openmessage CLAUDE.md?

Project instructions for OpenMessage, a local message database and built-in MCP server. OpenMessage imports conversations from Google Messages, Google Chat, iMessage, and WhatsApp into one system.

Developing, debugging, supporting, or extending OpenMessage's Go commands, SQLite storage, message importers, MCP tools, web interface, and macOS app.

They explain the project structure and operating rules so an AI assistant can work with imported messages and the local command-line and web applications correctly.

openmessage CLAUDE.md — Instructions file by MaxGhenis

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/maxghenis/openmessage/claude-md

Clone the repo

git clone --depth 1 https://github.com/MaxGhenis/openmessage

Per session 2,855 This file is loaded in full into every session.

When invoked 2,855 The same file — it is already loaded in full.

Security scan A 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.

Model	Per session	Once invoked
Fable 5	$0.02855	$0.02855
Opus 5	$0.01427	$0.01427
Sonnet 5	$0.00571	$0.00571
Haiku 4.5	$0.00285	$0.00285

Measured 2d ago against content hash 3289f271f71a, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

openmessage CLAUDE.md scanned grade A 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 2d 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.

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

curl -s -o /dev/null -w "%{http_code}" https://openmessage.ai

CLAUDE.md · 226 lines

How it starts

The opening of the file, as written. The whole thing — 226 lines — stays where its author put it; the contents beside it link to each section on GitHub.

OpenMessage

Local-first universal message database with built-in MCP server. Ingests messages from SMS/RCS (Google Messages), Google Chat, iMessage, and WhatsApp.

Architecture

├── cmd/              Go CLI commands (pair, serve, send, read, status, import)
├── internal/
│   ├── app/          Bootstrap, data dir, backfill
│   ├── client/       libgm Google Messages protocol
│   ├── db/           SQLite store (conversations, messages, contacts, unified_contacts, drafts)
│   ├── importer/     Multi-platform import adapters (gchat, imessage, whatsapp)
│   ├── story/        Stats computation + narrative story generation
│   ├── tools/        MCP tools (24 tools)
│   ├── viz/          Relationship visualization renderer (self-contained HTML)
│   └── web/          HTTP API + embedded React UI
├── macos/            Swift macOS app wrapper
│   ├── OpenMessage/  Swift package (BackendManager, PairingView, etc.)
│   └── build.sh      Builds universal binary + .app + .dmg
├── site/             Static website (deployed to openmessage.ai)
└── vercel.json       Vercel config (root — NOT site/vercel.json)

Supporting a live install (READ FIRST for support/debug tasks)

If you are debugging a real user's install — sends failing, re-pairing, reading their actual messages — read docs/agent-runbook.md before touching anything. The traps that cost the most:

Two data dirs, not one. The macOS app's live store is ~/Library/Application Support/OpenMessage/ (set via OPENMESSAGES_DATA_DIR). The CLI default (~/.local/share/openmessage/) is a separate, usually stale store. Point CLI tools at the app dir for live data.
Read live messages via the HTTP API (/api/conversations/<id>/messages, /api/search, /api/status) — the app holds the WAL'd DB, so a direct sqlite3 reader hits "unable to open database file (14)".
Re-pairing Google Messages: QR is dead; use Google Account pairing via the cookie method; clear session.json from both data dirs to reach the pairing screen; don't over-reconnect (it throttles the account). Full recipe in the runbook.

Read the full file on GitHub · 226 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.

2d ago First seen · 226 lines · 2,855 tokens per session scan A 3289f271f71a

Subscribe to this mod's changes

openmessage CLAUDE.md is an instructions file published in the GitHub repository MaxGhenis/openmessage (52 stars, last pushed 8d ago), licensed Unlicense. It adds 2,855 tokens to every session, about $0.0143 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

google-messages local-first macos mcp rcs

openmessage CLAUDE.md

Getting it into your agent

What it costs to keep this loaded

Grade A, and why

Makes network callslowCapability

How it starts

OpenMessage

Architecture

Supporting a live install (READ FIRST for support/debug tasks)

What this file has done since we first saw it

Other instructions, from other repositories

personal-model AGENTS.md

humla CLAUDE.md

os-clovy AGENTS.md

humla AGENTS.md

personal-model CLAUDE.md

Starcat AGENTS.md