Sandwalk
CLI

Doctor

Run environment, model provider, sandbox, and MCP readiness checks.

sandwalk doctor

In interactive mode, type /doctor (or ask Capy to run Doctor).

What sandwalk doctor Checks

  • Python & Core Scientific Packages: Verifies Python 3.11+, rdkit, duckdb, scanpy, and sandbox dependencies.
  • Model Providers & Authentication: Validates configured API keys, ChatGPT Plus/Pro OAuth tokens (~/.sandwalk/chatgpt-auth.json), AWS Bedrock credentials, and local llama.cpp / Ollama endpoints.
  • Skills, MCPs & Omnigraph: Checks installed skills in ~/.sandwalk/skills/, MCP server definitions in ~/.sandwalk/mcp.json, and the local Omnigraph binary (~/.sandwalk/bin/omnigraph).
  • Database & Storage Integrity: Verifies ~/.sandwalk/agent-server.db and ~/.sandwalk/memory/memory.db WAL health and disk space.

Last updated on

Data

Pull and manage local scientific reference corpora (DepMap, MSigDB, STRING).

Configuration

Everything under ~/.sandwalk/, environment variables, and LAN server authentication.

On this page

What sandwalk doctor Checks