shunt-codex

shunt-codex

Claude Code subagents that run on ChatGPT/Codex GPT-6 models — Sol and Luna — and GPT-5.6 models — Luna, Sol, and Terra — routed through the shunt gateway.

Unlike a CLI hand-off (which drops persona and preloaded skills), shunt diverts only token generation at the inference layer. The session keeps running inside Claude Code's harness: same tool loop, same skills, same script-path resolution. Only the model that generates the tokens changes.

Agents

Agent (@-mention)	Model id (model:)	Native effort	Supported effort levels
shunt-codex:gpt-6-sol	gpt-6-sol	medium	low · medium · high · xhigh · max · ultra
shunt-codex:gpt-6-luna	gpt-6-luna	medium	low · medium · high · xhigh · max
shunt-codex:gpt-5.6-sol	gpt-5.6-sol	low (fast)	low · medium · high · xhigh · max · ultra
shunt-codex:gpt-5.6-terra	gpt-5.6-terra	medium	low · medium · high · xhigh · max · ultra
shunt-codex:gpt-5.6-luna	gpt-5.6-luna	medium	low · medium · high · xhigh · max

Each agent's model: frontmatter pins the request to a Codex slug, so only that subagent diverts — the main session stays on Claude. The three GPT-5.6 agents share a 372k-token context window. GPT-6-Sol is the workhorse for coding and everyday work; GPT-6-Luna is the fast, affordable tier for easier tasks. gpt-6-sol and gpt-6-luna need a Codex client of at least 0.155.0 — shunt advertises one for you (see Prerequisites).

Effort levels are from openai/codex's models.json. Note the difference: Luna (both gpt-6-luna and gpt-5.6-luna) does not support the ultra level — its top level is max. Sending ultra to Luna (via shunt's effort config or an effort: override) is rejected upstream. Sol and Terra accept ultra.

The agents' system prompts are written for Claude Code's harness — these models run inside Claude Code's tool loop (Read/Edit/Bash, skills), not Codex's. The gpt-6-* and gpt-5.6-* entries in models.json do carry a substantial Codex system prompt (model_messages.instructions_template, ~16k chars, "You are Codex…"), but it describes Codex's own tools, personality, and workflow — none of which exist here. shunt diverts only token generation, not that prompt: Claude Code supplies its own system prompt, so the Codex one is never sent and there is nothing Codex-specific to reproduce in these agents.

Prerequisites

These agents only work when a shunt gateway is running in front of Claude Code and is configured to route the model ids above to the Codex provider:

1. Run shunt and point Claude Code at it:

   export ANTHROPIC_BASE_URL=http://127.0.0.1:3001   # shunt's default bind address

2. Authenticate the ChatGPT/Codex subscription shunt reuses:

   codex login   # writes ~/.codex/auth.json, which shunt reads + auto-refreshes

3. Map the slugs in your shunt.toml to the Codex provider (see shunt.toml.example):

   # [[routes]] is a TOML array-of-tables: one block per model slug.
   [[routes]]
   model = "gpt-6-sol"
   provider = "codex"
   
   [[routes]]
   model = "gpt-6-luna"
   provider = "codex"
   
   [[routes]]
   model = "gpt-5.6-sol"
   provider = "codex"
   
   [[routes]]
   model = "gpt-5.6-terra"
   provider = "codex"
   
   [[routes]]
   model = "gpt-5.6-luna"
   provider = "codex"

   For the full setup — auth-file handling, effort, context-window sizing, and troubleshooting — see the Codex configuration reference (site guide · docs/codex-configuration.md).

The ChatGPT-account backend only accepts the slugs your account is entitled to. The latest are gpt-6-astra / gpt-6-sol / gpt-6-luna / gpt-6.1-sol / gpt-6.1-terra, followed by gpt-5.6-sol / gpt-5.6-terra / gpt-5.6-luna; older accounts may only have gpt-5.5 / gpt-5.4. In models.json, gpt-6-sol and gpt-6-luna require a Codex client of at least 0.155.0, while gpt-6-astra, gpt-6.1-sol and gpt-6.1-terra require at least 0.153.0 (though gpt-6.1-sol was rejected with a 0.156.0 identity anyway); shunt's pinned client identity (codex_cli_rs/0.159.3) satisfies all five. This plugin ships no agent for gpt-6-astra, gpt-6.1-sol or gpt-6.1-terra, and a route alone does not make one invokable. To use one, add a [[routes]] block like those above, then pick one way to select it: ANTHROPIC_CUSTOM_MODEL_OPTION=gpt-6.1-sol to offer it in the /model picker, CLAUDE_CODE_SUBAGENT_MODEL=gpt-6.1-sol to run every subagent on it, or your own agent file — for example .claude/agents/gpt-6.1-sol.md, copied from this plugin's agents/gpt-6-sol.md with name and model: changed to gpt-6.1-sol and the GPT-6-Sol references in its description and prompt updated to match. The canonical catalog is openai/codex's models.json.

Without a running shunt gateway mapping these ids, Claude Code will send the gpt-6-* or gpt-5.6-* model id straight to Anthropic and the request will fail.

Install

/plugin marketplace add pleaseai/shunt
/plugin install shunt-codex@shunt

Usage

@shunt-codex:gpt-6-sol  refactor this module and run the tests

Or set every subagent to a Codex model for a session with CLAUDE_CODE_SUBAGENT_MODEL=gpt-6-sol.

Both require a running shunt gateway with the slug routed to the Codex provider — see Prerequisites. Without it the request fails against Anthropic.

Further reading

• Codex configuration reference — the full end-to-end setup (Markdown source).
• Effort & Context — reasoning effort, count_tokens, and the 372k context window in depth.
• Model Discovery — auto-list Codex models in the /model picker via a claude--named alias.
• ChatGPT / Codex auth spec — how shunt reads and refreshes ~/.codex/auth.json.

License

MIT OR Apache-2.0, matching the shunt project.