Documentation: Missing usage examples for Codex and Gemini agent types

Problem

The README prominently mentions that Axon supports three agent types:

• Claude Code
• OpenAI Codex
• Google Gemini

However, all documentation and examples only show Claude Code usage. There are no examples demonstrating how to use Codex or Gemini.

Current State

1. The README line 9 says: "runs AI coding agents (Claude Code, OpenAI Codex, Google Gemini)"

2. The README line 68 says: "Currently supported agents: Claude Code, OpenAI Codex, and Google Gemini."

3. All examples in the README use Claude Code:

   • Quick Start uses oauthToken (Claude Code specific)
   • All TaskSpawner examples use claude-code type
   • Config file template shows Claude credentials only

4. The agent implementations exist (codex/ and gemini/ directories with Dockerfiles and entrypoints)

5. The agent-image-interface.md doc mentions the environment variables (CODEX_API_KEY, GEMINI_API_KEY) but provides no usage examples

Why This is a Problem

1. Discovery gap: Users don't know how to actually use these agents
2. Credibility gap: Prominently advertising support for agents without showing how to use them makes the feature feel incomplete
3. Onboarding friction: Users who want to use Codex or Gemini have to reverse-engineer the configuration from code

What's Missing

Users need to know:

1. How to configure credentials for Codex/Gemini
2. Which model names are valid for each agent type
3. Any agent-specific differences or limitations
4. A complete working example for each agent type

Suggested Fix

Add a section to the README (or a separate doc) with examples for each agent type:

Example for Codex

# ~/.axon/config.yaml
apiKey: <your-openai-api-key>

# Or set CODEX_API_KEY directly:
# apiKey: sk-...

axon run -p "Add unit tests" --type codex --model gpt-4-turbo

Example for Gemini

# ~/.axon/config.yaml  
apiKey: <your-google-api-key>

axon run -p "Add unit tests" --type gemini --model gemini-2.0-flash

Additional Context

• The codex and gemini entrypoints exist and look production-ready
• The controller code supports these agent types
• This is just a documentation gap, not a feature gap

Impact

Without this documentation:

• Users may not realize Axon supports multiple agents
• Users who want to use non-Claude agents will struggle
• The multi-agent support feels like vaporware despite being fully implemented

Start with the README and agent-image-interface.md, then inspect the codex/ and gemini/ entrypoints to verify the documented credentials, model names, and agent-specific differences. Add complete Codex and Gemini configuration and command examples, and confirm that each example matches the existing agent implementations.