| name | help |
| description | Overview of all available GoodMem skills and what a user needs to get started. Invoke when a user asks what GoodMem skills are available, how to get started, or what they need to set up. |
GoodMem Plugin — Help & Overview
GoodMem is memory infrastructure for AI agents — store, retrieve, and manage memories with vector embeddings, semantic search, and RAG.
Credential Setup
Before calling any GoodMem MCP tool, you MUST ensure credentials are configured. Follow this sequence:
- Try calling
goodmem_system_info to check if credentials are already set (from environment variables).
- If it fails with a connection or auth error, ask the user for:
- Base URL — their GoodMem server address (e.g.,
https://your-server.example.com)
- API key — starts with
gm_
- Call
goodmem_configure with the provided base_url and api_key.
- Confirm the connection succeeded before proceeding.
Credentials persist for the entire session. The user can call goodmem_configure again to switch servers.
If the user has set GOODMEM_BASE_URL and GOODMEM_API_KEY environment variables before starting Codex, credentials are picked up automatically — no setup needed.
TLS errors? If goodmem_system_info fails with a certificate verification error (e.g., "self-signed certificate", "unable to verify the first certificate", or "CERT_HAS_EXPIRED"), the GoodMem server is using a self-signed or private CA certificate. Tell the user to set one of these environment variables before launching Codex:
NODE_EXTRA_CA_CERTS=/path/to/rootCA.pem — adds the CA to the trusted store (recommended)
NODE_TLS_REJECT_UNAUTHORIZED=0 — disables certificate verification (local dev only)
Available Skills
mcp — MCP Tools Reference
Reference for all GoodMem MCP tools. Use this when you need to look up tool names, parameters, or behavior. The MCP tools let you operate GoodMem directly — create embedders, store memories, run retrieval — all via natural language.
python — Python SDK Reference
Complete API reference for the goodmem Python package. Use this when writing Python code that integrates with GoodMem. Includes method signatures, convenience shortcuts, model identifiers, and code examples.
help — This skill
Overview and setup instructions.
Two Ways to Use GoodMem
1. Direct operations (MCP tools)
Ask Codex to perform GoodMem operations in natural language. Codex calls MCP tools behind the scenes.
Examples:
- "Create an embedder called 'my-embedder' for the model I want to use"
- "Create a space called 'knowledge-base' using that embedder"
- "Store this text as a memory in the knowledge-base space"
- "Search my knowledge-base for information about authentication"
- "List all my spaces"
2. Python code generation (SDK skill)
Ask Codex to write Python code using the GoodMem SDK.
Examples:
- "Write a Python script that creates an embedder and stores some documents"
- "Show me how to do semantic retrieval with the GoodMem SDK"
- "Write a RAG pipeline using GoodMem for memory storage"
Common Workflows
Quick start with MCP
- "Create an embedder for the model I want, using the provider credentials I have"
- "Create a space called 'docs' using that embedder"
- "Store this text as a memory: [your content]"
- "Search docs for [your query]"
Build a RAG app with the SDK
- "Write a Python script that sets up a GoodMem RAG pipeline"
- Codex generates code using the
goodmem Python package
- Run the generated script