agentmash 0.3.0

package/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 AgentMash contributors
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

package/README.md

@@ -0,0 +1,72 @@
+# agentmash
+
+Let your team's Claude Code agents see each other's work.
+
+When several people run Claude Code against the same repository, their agents collide: two
+of them edit the same file, one breaks the interface the other is building against, both
+write the same helper. AgentMash fixes that with Claude Code hooks — no fork, no wrapper,
+no MCP server.
+
+## Setup
+
+In your team's repo, once:
+
+```bash
+npx agentmash init
+```
+
+That creates a room on the coordination server, installs the hook scripts into
+`.claude/agentmash/`, registers them in `.claude/settings.json` without disturbing anything
+already there, writes the room config beside them, and prints your dashboard link. Commit
+it:
+
+```bash
+git add .claude && git commit -m "add agentmash" && git push
+```
+
+Your teammates run `git pull`. That is the whole setup for them — the hooks and the room
+travel with the repo, so their next Claude Code session joins automatically. Nothing to
+install, no token to paste.
+
+Node 18 or newer. The CLI has no dependencies and needs no checkout of this project.
+
+## Commands
+
+| Command | What it does |
+| --- | --- |
+| `npx agentmash init` | Create a room and wire up this repo |
+| `npx agentmash join <id\|url>` | Point this repo at an existing room |
+| `npx agentmash doctor` | Check exactly what is and isn't working |
+| `npx agentmash status` | Who is working on what, from the terminal |
+| `npx agentmash uninstall` | Remove the hooks from this repo |
+
+Flags: `--server <url>` (coordination server), `--repo <path>`, `--label <text>`,
+and `--force` with `init` to replace an existing room.
+
+Reach for `doctor` when someone's events aren't showing up. The hooks fail silently by
+design, so it is the only thing that will tell you why: it resolves the config and shows
+where each value came from, checks the hooks are registered, measures the round trip and
+compares it against what a cold hook needs, validates the room, and lists the teammates it
+can see.
+
+## Running your own server
+
+`init` points at a shared server by default. To coordinate on your own infrastructure,
+deploy the server from the repository (Fly.io and Docker Compose configs are included) and
+point repos at it:
+
+```bash
+npx agentmash init --server https://my-agentmash.example.com
+```
+
+## What leaves your machine
+
+File paths, branch names, and the first ~200 characters of each prompt, so teammates can
+see what you are working on. No file contents, ever. The room id in `.claude/agentmash/`
+is the credential, so repo access is the security boundary — use a private repo, or
+`agentmash join` a room shared out of band.
+
+Full documentation, self-hosting instructions and limitations:
+<https://github.com/dauletbekalim/agentmash>
+
+MIT licensed.