Siphon — Development skill for Claude Code
Siphon: a raw-C, zero-dependency GGUF inference engine for Linux.
How to install Siphon
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open FemBoyGamerTechGuy/Siphon and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Siphon does
Siphon: a raw-C, zero-dependency GGUF inference engine for Linux. Runs giant models on modest hardware by splitting work across GPU (Vulkan, no CUDA), CPU and disk. Four resource modes, on-disk KV cache, and OpenAI/Anthropic-compatible API for agent tools like Claude Code and opencode.
Alternatives in Development
- Happy Coder — by GrocerPublishAgent - Spawn and control multiple Claude Codes in parallel from your phone or desktop 15.7k ★
- Microsandbox — Open-source self-hosted MicroVM sandboxes with sub-200ms startup, hardware-level isolation via libkrun 5k ★
- Investigate Performance — Investigate a Nimbalyst performance problem (freeze, lag, idle CPU, slow op, memory growth) with measured evid 1.6k ★
README
Siphon
Siphon: a raw-C, zero-dependency GGUF inference engine for Linux. Runs giant models on modest hardware by splitting work across GPU (Vulkan, no CUDA), CPU and disk. Four resource modes, on-disk KV cache, and OpenAI/Anthropic-compatible API for agent tools like Claude Code and opencode.
Related Skills
GPU Check
Check GPU availability + driver version — ROCm untuk AMD RX 6700S, CUDA untuk NVIDIA GTX 1650.
Magnitude
Open source inference server your agent sets up for you. It profiles your hardware, helps you choose the best
Herdr Agent Resource Badges
herdr plugin: green/orange/red CPU and RAM badges per agent in the sidebar, shown only while an agent is busy.
Unity Profile
Deep profiling session — captures frames via MCP, analyzes CPU/GPU timing, memory snapshots, rendering stats,
StrataGP
Qwen3.8-Flash-Next (125B MoE) on a 8GB+ NVIDIA GPU: one-click install for Windows / Linux. Strata inference en
Deeplit MCP Server
Claude Code and Codex plugin for deeplit: manage GPU inference instances, volumes, models and billing from you
Related Agents
Qe
Performs functional and acceptance testing against the task spec — Julia Pkg.test (CPU always, CUDA/GPU when r
Lemonade Specialist
Lemonade Server and SDK specialist for local LLM on AMD hardware. Use PROACTIVELY for Lemonade setup, model ma
Inference Architect
You are the system Inference Architect — a senior ML infrastructure engineer specializing in local LLM serving
