llnl / llnl/mada

Per-Agent Model Override

Open
#45 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

enhancement
Dominant language
Python
Stars
2
Forks
2
Avg merge
1d 4h
Merged PRs (30d)
4

Description

We should have a way for users to specify different models for each agent. I think the functionality should be to have a default model that all agents use unless a user specifically provides a different one.

Config should likely change to something similar to this:

{
  "models": {
    "default": {  // This 'default' model is required in the config
      "model": "gpt-5.4",
      "api_key": "${API_KEY}",
      "base_url": "${API_BASE_URL:-https://livai-api.llnl.gov/v1}"
    },
    "my_model": {
      "model": "claude-sonnet-3.7",
      "api_key": "${API_KEY}",
      "base_url": "${API_BASE_URL:-https://livai-api.llnl.gov/v1}"
    }
  },
  "agents": [
    {
      "agent_name": "AgentWithDefaultModel",
      "description": "Agent using the default model (no 'model' setting)",
      "domain": "my_domain",
      "mcp_servers": ["some_server"],
      "instructions": "Do something",
    },
    {
      "agent_name": "AgentWithCustomModel",
      "description": "Agent using a custom model (has 'model' setting)",
      "domain": "my_domain",
      "mcp_servers": ["some_server"],
      "instructions": "Do something",
      "model": "my_model"
    }
  ]
}

Contributor guide

No contributing guide indexed for this repository

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

No files or tests are named. Start by locating the configuration parser and agent/model construction entry points, then trace how the current model is selected. Done means a required default model works for agents without an override and an explicitly named model is used for agents that specify one.

Written by the indexing model from the issue text.

Assessment

Tech stack
python
Domain
ai, backend
Issue type
Feature
Difficulty
4/5
Estimated time
3-5 days
Activity status
Quiet
Clarity
Mostly clear
Newbie friendliness
62/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.