rjv-codex-ollama-subagentslisted

# Codex + local Ollama Two ways to put a local Ollama model to work — pick by whether the job needs the repo: ```text (a) Codex native subagents WITH repo tools → most work; the bulk of this skill main Codex GPT-5.5 = orchestrator / reviewer qwen-explorer / gemma-explorer = local read-only repo scouts qwen-worker / gemma-worker = local scoped mechanical edit workers optional OpenAI mini = fast cheap explorer when latency matters (b) One-shot local text function, NO repo tools → see the last section complete prompt in → local Ollama text out (draft-from-spec, tests, classify) ``` **Which mode:** if the model must read files or edit through the harness → (a), Codex subagents. If you can hand it a complete self-contained prompt and it needs nothing else from the repo → (b), the one-shot runner. The tell you picked wrong for (b): you wish it could "check the other file" — that's a subagent job, route it to (a). ## Optional install: hybrid Codex profile Only run the installer when the user explicitly wants this machine configured. Do not run it just because the skill loaded. Prereqs: - `codex` installed - `ollama` installed - local models available, for example `qwen3:8b`, `qwen3.6:35b`, and `gemma4:26b` From this skill directory: ```sh scripts/install-codex-ollama-profile.sh ``` Reruns are safe: existing files are skipped. To update existing profile/agent files from this skill, run: ```sh FORCE=1 scripts/install-codex-ollama-profile