harivansh-afk/sandbox-agent
1
0
Fork 0
mirror of https://github.com/harivansh-afk/sandbox-agent.git synced 2026-04-15 07:04:48 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

feat: refresh docs and agent schema

Browse source
This commit is contained in:
Nathan Flurry 2026-01-25 03:04:12 -08:00
parent a49ea094f3
commit 0fbf6272b1
39 changed files with 3127 additions and 1806 deletions
Download patch file Download diff file Expand all files Collapse all files

95
README.md
View file

@ -1,94 +1,15 @@
# Sandbox Agent SDK
Run inside sandboxes to provide support
Universal API for running Claude Code, Codex, OpenCode, and Amp inside sandboxes.
- **Any coding agent**: Universal API to interact with all agents with full feature coverage
- **Server Mode**: Run as HTTP server from any sandbox provider or as TypeScript & Python SDK
- **Universal session schema**: Universal schema to store agent transcripts
- **Supports your sandbox provider**: Daytona, E2B, Vercel Sandboxes, [add your own](TODO)
- **Lightweight, portable Rust binary**: Install anywhere with 1 curl command
- **Compatible with Vercel AI SDK**: TODO
Documentation lives in `docs/` (Mintlify). Start with:
## Quickstart
- `docs/index.mdx` for the overview
- `docs/quickstart.mdx` to run the daemon
- `docs/http-api.mdx` and `docs/cli.mdx` for API references
Start with the SDK:
Quickstart (local dev):
```bash
sandbox-agent --token "$SANDBOX_TOKEN" --host 127.0.0.1 --port 8787
```
TODO
```
To run this in server mode, install with:
```
TODO
```
And run with:
```
TODO
```
See the example for your provider of choice:
- TODO
- [Add your own](TODO)
## Security
TODO: Tokens
TODO: Using a gateawy
TODO: BYO tokens with extractor
## Demo Frontend
TODO: Screenshot
This project provides a demo frontend for testing the connection. Run it with:
```
TODO
```
## Agent Compatibility Matrix
TODO
## Reference
### TypeScript SDK
TODO
### HTTP API
TODO
### CLI
TODO
## FAQ
TODO
- Why not use PTY? This is the recommended option for XXXX
- Why not use <feature that already exists on sandbox API>?
- Does it support <platform>?
- Can I use this with my personal OpenAPI & Claude tokens?
## Project Scope
This project aims to solve 3 problems with agents:
- **Universal Coding Agent API**: Claude Code, Codex, Amp, and OpenCode all have put a lot of work in to the agent scaffold. Each have respective pros and cons and need to be easy to be swapped between.
- **Agents In Sandboxes**: There are many complications with running agents inside of sandbox providers. This lets you run a simple curl command to spawn an HTTP server for using any agent from within the sandbox.
- **Agent Transcript**: Maintaining agent transcripts is difficult since the agent manages its own sessions. This provides a simpler way to read and retrieve agent transcripts in your system.
Features out of scope:
- **Storage of sessions on disk**: Sessions are already stored by the respective coding agents on disk. It's assumed that the consumer is streaming data from this machine to an extral storage, such as Postgres, ClickHouse, or Rivet.
- **Direct LLM wrappers**: Use the [Vercel AI SDK](https://ai-sdk.dev/docs/introduction) if you want to impelment your own agent from scratch
- **Git Repo Management**: Just use git commands or the features provided by your sandbox provider of choice.
- **Sandbox Provider API**: Sandbox providers have many nuanced differences in their API, it does not make sense for us to try to provide a custom layer. Instead, we opt to provide skills that lets you integrate this project with sandbox providers.