Skip to content

Security: aralde/astorLM

Security

SECURITY.md

Security Policy

Supported versions

astorlm is pre-1.0. Security fixes land on main and ship in the next release; older minor versions are not backported.

Version Supported
0.x (latest) ✅
older 0.x ❌

Reporting a vulnerability

Please do not open a public issue for security reports.

Use GitHub's private reporting instead: Report a vulnerability.

Include, as far as you can:

  • the affected version, entrypoint and platform,
  • a minimal reproduction,
  • the impact you believe it has.

You can expect an acknowledgement within a few days. Once a fix is ready, it is released and the advisory is published with credit to you, unless you prefer to stay anonymous.

Threat model — what this library does and does not protect

astorlm is a library that lets a language model run tools on your behalf. Understanding the boundaries matters more than any single report:

  • The default Executor runs commands on the host. LocalExecutor gives the model the same privileges as the process embedding the library. It is a convenience backend, not a sandbox. If the model's input is untrusted, use DockerExecutor (OS-level isolation) or a custom Executor of your own.
  • resolveSafe constrains paths to cwd, not the shell. The built-in read/write/edit tools refuse to escape the working directory, but the bash family executes whatever command it is given. Path confinement is not command confinement.
  • QuickJsCodeRunner is the sandboxed tier. It runs self-contained source inside a memory-safe WASM runtime with no filesystem, network or host syscalls unless you grant them explicitly.
  • Prompt injection is a real risk for any agent. Content the model reads — files, web pages, MCP tool results — can contain instructions. The beforeToolExecution hook is the intended enforcement point: use it to require authorization for sensitive tools. See restrictToolsHook for a ready-made allowlist.
  • MCP servers are trusted code. mountMcpServer runs the transport you configure and does not validate tool inputs — the server is expected to. Only mount servers you trust.
  • Credentials. API keys are read from the environment through AuthStorage and passed to providers. The library does not persist them. Session files written by FileSessionManager contain conversation content, which may include whatever the model or your tools put there — treat session directories as sensitive.

Reports about these documented behaviours are not vulnerabilities in themselves, but reports showing a way to bypass a boundary the library claims to enforce definitely are.

There aren't any published security advisories