ouroboros-config

# /ouroboros:ouroboros-config Edit `~/.ouroboros/config.yaml`: global and per-stage runtime and model choices, the LLM backend, install badges for missing CLIs, and env-override warnings. ``` ooo config /ouroboros:ouroboros-config ``` ## Instructions Pick the branch by one question: **can the user open a browser pointed at this machine?** Run commands with the active runtime's `run_shell` capability. ### Branch A: local harness on the user's own machine 1. Launch in the background (it serves until stopped): ```bash if command -v ouroboros >/dev/null 2>&1; then ouroboros config else uvx --python '>=3.12' --from 'ouroboros-ai[tui]' ouroboros config fi ``` In a non-interactive context it serves the settings app on a local web server and opens the browser. The `uvx` fallback covers a plugin-only install where `ouroboros` is not on `PATH`; in a development checkout use `uv run ouroboros config`. 2. Relay the `http://localhost:<port>` line in case the browser did not open. 3. Tell the user: edit, Save, then ask you to stop the server. A running MCP server may need a reconnect to pick up backend changes. `ooo config` reopens the settings any time; a saved model choice is never locked. ### Branch B: remote host the user can reach (SSH box, home server) ```bash ouroboros config --web --host 0.0.0.0 --no-browser ``` Relay the printed URL with this host's address substituted, plus the SSH tunnel fallback it prints (`ssh -L <port>:localhost:<port> <this-host>`). ### Branch C: chat gateway with no browser path (e.g. hermes driven from Discord) Do not start a server nobody can reach. Drive the same settings in chat: 1. Show the current state with `ouroboros config show`. 2. Offer a short menu with current values (default agent, per-stage agents, and models: `auto`, a tier, or a model id for every role) and ask what to change. 3. Apply each choice with the validated setter (same write path as the GUI): ```bash ouroboros config set orchestrator.runtime_backend <agent> ouroboros config set orchestrator.runtime_profile.stages.<interview|execute|evaluate|reflect> <agent> ouroboros config set models.default <auto|frugal|standard|frontier|model> # every role ouroboros config set models.pin true # run the per-role ids below ouroboros config set clarification.default_model <model> # interview & seed ouroboros config set execution.default_model <model> # execute ouroboros config set evaluation.semantic_model <model> # evaluate ouroboros config set resilience.reflect_model <model> # reflect ouroboros config set llm.backend <backend> # internal LLM calls ``` 4. Confirm with `ouroboros config show` and summarize what changed. ### All branches Relay any rejected `set` validation error or missing-dependency hint (`pip install 'ouroboros-ai[tui]'`) verbatim. `ouroboros config undo` restores the configuration saved before the last change. ## RFC #1392 State Breadcrumb Footer End the final response with exactly one line `◆ <current state> → next: <recommended action>`. Derive the state from live session state via `ouroboros_session_status` when available, otherwise from this skill's actual outcome. Never use a linear `Step N of M` footer; when the next step is a real choice, list 2-3 options.