exoagent 0.0.1

package/LICENSE.md

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2025-PRESENT Ryan Rasti <https://github.com/ryanrasti>
+
+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 +1,66 @@
+# exoagent
+
+The OS kernel to safely unleash your agents.
+
+**[Try the challenge](https://exoagent.io/challenge)** — Two agents, same LLM, same prompt injection vulnerability. $1,000 in BTC if you can hack the one protected by ExoAgent (coming soon!).
+
+## The Problem
+
+Today's agent frameworks give LLMs raw access to tools. The "security model" is hoping the prompt works.
+
+- **Authorization is broken** — Tool calls inherit user permissions. Your agent gets your credentials — all of them.
+- **Interfaces are opaque** — `execute_sql("SELECT * FROM users")` — policy can't see what's inside.
+- **No central data policy** — Each tool enforces its own rules. No holistic view. No real guarantees.
+
+## The Fix
+
+Security as a system invariant, not a polite suggestion.
+
+- **Object-capability tools** — Instead of flat tools, pass your agents dynamic objects with strict security guarantees
+- **Semantic interfaces** — Rich, secure interaction starting with SQL
+- **Central policy** — Declare & enforce what data can flow where
+
+## Installation
+
+```bash
+npm install exoagent
+```
+
+## Quick Start
+
+```ts
+import { Database, RpcToolset, tool } from 'exoagent'
+
+class User extends db.Table('users').as('user') {
+  id = this.column('id')
+  name = this.column('name')
+  email = this.column('email')
+
+  @tool()
+  orders() {
+    return Order.on(order => order.userId['='](this.id)).from()
+  }
+}
+
+// Agent can only access what you've exposed
+class MyAgent extends RpcToolset {
+  @tool()
+  users() {
+    return User.on(user => user.id['='](this.currentUserId)).from()
+  }
+}
+```
+
+## Roadmap
+
+- [ ] Additional SQL support
+- [ ] Policy engine — information flow control to declare & enforce what data can flow where
+- [ ] Python port
+
+## Documentation
+
+Coming soon.
+
+## License
+
+[MIT](./LICENSE) License © [Ryan Rasti](https://github.com/ryanrasti)