From llm-externalizer
Use when inspecting LLM Externalizer profile config or explaining the manual-edit policy. Trigger with "show LLM profile", "which model is active", "edit settings.yaml".
npx claudepluginhub emasoft/emasoft-plugins --plugin llm-externalizer[list]This skill uses the workspace's default tool permissions.
Profile-based configuration for local and remote LLM backends. Settings stored at `~/.llm-externalizer/settings.yaml`. Three modes: `local` (sequential), `remote` (parallel via OpenRouter), `remote-ensemble` (three models in parallel).
Guides Next.js Cache Components and Partial Prerendering (PPR): 'use cache' directives, cacheLife(), cacheTag(), revalidateTag() for caching, invalidation, static/dynamic optimization. Auto-activates on cacheComponents: true.
Processes PDFs: extracts text/tables/images, merges/splits/rotates pages, adds watermarks, creates/fills forms, encrypts/decrypts, OCRs scans. Activates on PDF mentions or output requests.
Share bugs, ideas, or general feedback.
Profile-based configuration for local and remote LLM backends. Settings stored at ~/.llm-externalizer/settings.yaml. Three modes: local (sequential), remote (parallel via OpenRouter), remote-ensemble (three models in parallel).
Changing models, profiles, API keys, URLs, or timeouts is user-only. This is a deliberate design choice.
The MCP tools set_settings and change_model are disabled; the CLI subcommands profile add | select | edit | remove | rename are disabled. Calling any of them returns a refusal pointing here.
The only supported workflow is:
~/.llm-externalizer/settings.yaml in your editor.mcp__llm-externalizer__reset tool to reload without restarting.OPENROUTER_API_KEY environment variable setCopy this checklist and track your progress:
mcp__llm-externalizer__discover to report active profile, mode, model, auth status, service health, context window, concurrency limitsmcp__llm-externalizer__get_settings — returns a file path to an editable copy of settings.yaml~/.llm-externalizer/settings.yaml in an editor, then restarting or calling resetDo NOT attempt to call set_settings or change_model — both are disabled.
Use this skill when the user wants to see the current LLM Externalizer configuration or asks how to change a profile, a model, an API key, or a timeout. Always route configuration changes back to the user (manual YAML edit).
discover — active profile name, mode, model, auth status, context window, concurrency limitsget_settings — file path to a read-only editable copy (you still show the user how to apply changes manually)reset — reloads settings after the user has edited the file| Error | Cause | Resolution |
|---|---|---|
discover shows invalid profile | Settings file malformed | Ask user to open ~/.llm-externalizer/settings.yaml, fix, save, then call reset |
Auth token (NOT SET) | Env var missing | Ask user to add the env var to shell profile or .mcp.json, then restart Claude Code |
| Connection refused | Local LLM server not running | Ask user to start LM Studio / Ollama / vLLM; verify URL and port in settings |
set_settings / change_model returned DISABLED | Expected — tools are user-only | Direct user to manual edit |
1. discover → shows active: remote-ensemble-geminigrok, mode: remote-ensemble
2. get_settings → /path/to/settings_edit.yaml
3. Read the file, surface the profiles table to the user
4. Tell the user: to change, edit ~/.llm-externalizer/settings.yaml, save, then call reset
Do NOT call set_settings. Instead:
1. Show the current active profile via discover
2. List available profile names from get_settings
3. Tell the user: open ~/.llm-externalizer/settings.yaml, change the "active:" line to the desired profile name, save, then call reset
/llm-externalizer:llm-externalizer-configure — read-only inspector command/llm-externalizer:llm-externalizer-discover — quick health check