agentleFS
Sign inStart free
RoentekRoentek/Claude_Code_Boilerplate_Framework/.claude/skills/llmfit/SKILL.md

llmfit

Skill4 starsChanged 6 months agoMITView on GitHub

Right-size LLM models to your hardware. Detects GPU/CPU/RAM, scores every model for fit/speed/quality, downloads GGUF weights, runs inference. CLI-first; MCP (llmfit serve --mcp) as fallback. Source: https://github.com/AlexsJones/llmfit Use when you need to pipe results into other tools or maintain state across calls. MCP server runs via stdio: llmfit serve --mcp Configured in .mcp.json as llmfit-mcp. Use llmfit-mcp MCP tools when: - CLI output needs to feed into another MCP tool in the same turn - You need persistent server state…

  • Reads credentials

The file

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use when you need to pipe results into other tools or maintain state across calls.

    line 69

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

llmfit

Right-size LLM models to your hardware. Detects GPU/CPU/RAM, scores every model for fit/speed/quality, downloads GGUF weights, runs inference. CLI-first; MCP (llmfit serve --mcp) as fallback.

Source: https://github.com/AlexsJones/llmfit

When to use

  • "What LLM can run on my machine?"
  • "Compare llama3 vs mistral for my GPU"
  • "Download a GGUF model"
  • "Recommend a model with tool-use support"
  • "Plan hardware upgrade for a specific model"
  • "Benchmark my inference setup"

CLI (Primary — zero context tokens)

# Hardware detection
llmfit system --json

# Find models that fit your system
llmfit fit --json
llmfit fit --json --tool-use        # only models with function-call support
llmfit fit --json -p                # perfect-fit only (no compromise)
llmfit fit --json -n 10             # top 10

# Recommend top models for your hardware
llmfit recommend --json

# Search for a specific model
llmfit search "llama 3" --json
llmfit info "llama-3.3-70b" --json

# Compare two models side-by-side
llmfit diff "llama-3.3-70b" "mistral-7b" --json

# Plan hardware for a model you want
llmfit plan "llama-3.3-70b" --json

# Download GGUF weights from HuggingFace
llmfit download "model-name"

# Search HuggingFace for GGUF models
llmfit hf-search "llama 3 8b" --json

# Run a downloaded model
llmfit run "model-name"

# Benchmark inference against running providers
llmfit bench --json

# Update model database cache
llmfit update

Global flags

FlagEffect
--jsonStructured JSON output (for agent integration)
--memory <SIZE>Override GPU VRAM (e.g. "32G")
--ram <SIZE>Override system RAM
--cpu-cores <N>Override CPU core count
--max-context NCap context length for memory estimation

MCP (Fallback — structured in-session results)

Use when you need to pipe results into other tools or maintain state across calls.

MCP server runs via stdio: llmfit serve --mcp

Configured in .mcp.json as llmfit-mcp. Use llmfit-mcp MCP tools when:

  • CLI output needs to feed into another MCP tool in the same turn
  • You need persistent server state during a session

REST API (Optional)

# Start REST API + web dashboard at http://localhost:8787
llmfit serve --port 8787

Endpoints documented in API.md inside the llmfit repo.

Environment variables

VarPurpose
LOCALMAXXING_API_KEYCommunity benchmark data from localmaxxing.com (optional)
OLLAMA_CONTEXT_LENGTHDefault context cap when --max-context not set

Common patterns

# Find all tool-use capable models that fit exactly, output JSON
llmfit fit --json --tool-use -p

# Evaluate what would fit on a machine with 24GB VRAM
llmfit fit --json --memory 24G

# Download best-fit GGUF and run it
llmfit recommend --json | jq -r '.recommendations[0].name' | xargs llmfit download

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free