config

Inspect the local model configuration across all peer engines, show what is currently available, and offer targeted edits to the defaults and the per user capability scores. Treat this as the write path counterpart to /fusion:doctor. Steps: 1. Gather current state. - Capture environment overrides at invocation: - CLAUDE_CODE_SUBAGENT_MODEL: !`sh -c 'printenv CLAUDE_CODE_SUBAGENT_MODEL || echo "(unset)"'` - ANTHROPIC_MODEL: !`sh -c 'printenv ANTHROPIC_MODEL || echo "(unset)"'` - ANTHROPIC_DEFAULT_HAIKU_MODEL: !`sh -c 'printenv ANTHROPIC_DEFAULT_HAIKU_MODEL || echo "(unset)"'` - ANTHROPIC_DEFAULT_FABLE_MODEL: !`sh -c 'printenv ANTHROPIC_DEFAULT_FABLE_MODEL || echo "(unset)"'` - Read `~/.claude/settings.json` if it exists and report only the `model` key for the main session model. - Read only the frontmatter of every file in `${CLAUDE_PLUGIN_ROOT}/agents/` and report each Claude worker pin's `model` and `effort` values. - Run `grok models` to enumerate the account's live model lineup and current default, then extract only model related keys from `~/.grok/config.toml`. - Extract only the `model` and `model_reasoning_effort` keys from `~/.codex/config.toml`. - Run `node "${CLAUDE_PLUGIN_ROOT}/scripts/fusion-config.mjs" show` to display the per user quantified engine capability table and its backing file path. Use `node "${CLAUDE_PLUGIN_ROOT}/scripts/fusion-config.mjs" show --json` only when structured output makes the next edit decision clea...