From llm-externalizer
Use when scanning a project for free using the Nemotron model (no cost, lower quality). Trigger with "free scan", "free-scan", "scan for free", "quick scan", "cheap scan", "scan without cost", "nemotron scan".
npx claudepluginhub emasoft/emasoft-plugins --plugin llm-externalizer[folder-path | .]This skill uses the workspace's default tool permissions.
Run a project scan using the **free** NVIDIA Nemotron 3 Super model (`nvidia/nemotron-3-super-120b-a12b:free`). No cost, single model, 262K context.
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.
Run a project scan using the free NVIDIA Nemotron 3 Super model (nvidia/nemotron-3-super-120b-a12b:free). No cost, single model, 262K context.
LOW QUALITY: This model has significantly lower intelligence than the 3-model ensemble. Expect more false positives, missed bugs, and shallow analysis. Use only for quick rough checks on non-critical code.
WARNING: Prompts are logged by the provider — do not use with sensitive or proprietary code.
$OPENROUTER_API_KEY)Copy this checklist and track your progress:
mcp__plugin_llm-externalizer_llm-externalizer__discover to verify service is online/), or use current working directory.ts, .py → pass as extensionsexclude_dirsmcp__plugin_llm-externalizer_llm-externalizer__scan_folder with:{
"folder_path": "<parsed path or cwd>",
"free": true,
"use_gitignore": true,
"extensions": "<if parsed from prompt>",
"exclude_dirs": "<if parsed from prompt>",
"instructions": "<parsed instructions or default: 'Report REAL bugs only — logic bugs, crashes, security vulnerabilities with exploit paths, data corruption, functionality mismatch. DO NOT report missing error handling / null checks / validation / logging (fail-fast is a valid style choice). Respect the source file coding style. Reference function names + line numbers. Be terse.'>"
}
.md files EXCLUDED by default — source-code rubric is wrong for prose. Pass instructions describing a semantic search to include them. Use CPV or claude plugin validate . for structural validation.check_against_specs. For "already implemented?" hunts use search_existing_implementations.One .md report per source file, saved in reports_dev/llm_externalizer/. Each report contains findings from the single free model. Report filenames include the source filename for easy identification.
| Error | Cause | Resolution |
|---|---|---|
| Service offline | MCP server not running | Restart Claude Code or run /llm-externalizer:llm-externalizer-discover |
| Auth error | $OPENROUTER_API_KEY not set | Set the env var in .claude/settings.json or shell profile |
| Empty response | Model timed out | May need simpler instructions or smaller files |
| No files found | Wrong path or all files gitignored | Check folder_path is correct and use_gitignore setting |
{"tool": "scan_folder", "folder_path": "/path/to/project/src",
"free": true, "instructions": "Find security issues.",
"use_gitignore": true}
{"tool": "scan_folder", "folder_path": "/path/to/project",
"free": true, "extensions": [".py"],
"instructions": "Find TODO comments and classify by urgency."}