agentleFS
thenavidmthenavidm/whatsapp-mcp/AGENTS.md

whatsapp-mcp

AGENTS.md2 starsChanged 2 months agoMITView on GitHub

For agents editing this repository. Users read the README. A Go binary that links to WhatsApp as a companion device, the same protocol WhatsApp Web uses, and exposes the account over MCP. It is distributed through npm as a platform-specific wrapper, so a reader installs it with npx and never sees Go. Commit as n@navid.me. Never pass -c user.email=. The global config is correct and the override is the bug: the wrong address credits a blocked account and the Contributors…

  • Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.

    Of the AGENTS.md files we read, the 31,774 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.

  • In this repositoryOne other file it gives its agents.

The file

We split this AGENTS.md into the rules it gives an agent

1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Commit as n@navid.me: Never pass -c user.email=.

    line 25

≈ “Common” is an estimate: at least 1 in 5 of the 40,246 AGENTS.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
AGENTS.md

Working on whatsapp-mcp

For agents editing this repository. Users read the README.

What this is

A Go binary that links to WhatsApp as a companion device, the same protocol WhatsApp Web uses, and exposes the account over MCP. It is distributed through npm as a platform-specific wrapper, so a reader installs it with npx and never sees Go.

cmd/whatsapp-mcp/     entry point
internal/wa/          the WhatsApp client and media handling
internal/mcpserver/   tool registration
internal/safety/      the send guard: allowlist, rate limit, audit
internal/store/       local message history
internal/transcribe/  voice note transcription, optional
internal/agent/       the agent-facing surface
npm/                  the wrapper published to npm

Non-negotiables

Commit as n@navid.me. Never pass -c user.email=. The global config is correct and the override is the bug: the wrong address credits a blocked account and the Contributors panel reads 0.

Sending is OFF by default, and this repo is the exception to the house rule.

Everywhere else, writes work by default, because publishing is the point of a publishing tool. That reasoning does not transfer here:

  • The point of this server is reading your own history. Sending is a secondary convenience, not the reason anybody installs it.
  • A WhatsApp inbox is the most injectable surface in this whole family of servers. Every message is text a stranger chose, and a message in a group can instruct a model to forward private history somewhere.
  • A wrong send goes to a named human being who knows you, cannot be unsent, and costs a relationship rather than a deleted post.
  • WhatsApp bans accounts for automated behavior. A wrong send risks the number.

So AllowSend must be explicitly enabled, an allowlist can narrow sends to specific chats, sends are capped per rolling minute, and every attempt is appended to an audit log. Do not "correct" this to match the other servers.

The guard is the defence, not the README. Naming prompt injection in prose does nothing. internal/safety/guard.go is what actually stops it.

Never add anything that sends in bulk or on a schedule. Volume is what gets a number banned, and this is somebody's personal account.

Before claiming it works

go build ./... && go test ./...

A green suite is not a working server. Link a test device and run the real handshake before saying it works.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.