qwen-ollama

스타6

포크0

업데이트2026년 7월 27일 07:09

Use when a task needs local LLM inference via the on-box Ollama server (Qwen 2.5) instead of a cloud API.

설치

Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.

Manus에서 실행

출처

lawless-m

lawless-m/claude-skills

GitHub 저장소 열기 Creator 저장소 보기

다운로드

Manus에서 실행

Qwen via Ollama

Chosen model: qwen2.5:7b (4.7 GB). Other sizes exist if needed: 0.5b / 14b / 32b.
Endpoint: http://localhost:11434 (standard Ollama API).
Timeouts: 120 s default for analysis calls; longer (e.g. 300 s) for heavy generation.
Prefer local over cloud when: offline, data can't leave the machine, or bulk work where API cost matters.
VRAM contention with other GPU services: see the Vram-GPU-OOM-memory-management skill.

Reference implementation (Marvinous)

OllamaClient.rs in this folder — copy of the production async client (reqwest, timeout, error handling) from /home/matt/Marvinous/src/llm/client.rs.
Prompt building: /home/matt/Marvinous/src/llm/prompt.rs
System prompt (Marvin's personality): /etc/marvinous/system-prompt.txt
Config (model/endpoint settings): /etc/marvinous/marvinous.toml

이 저장소의 다른 Skills

같은 저장소

anthropic-rate-limits

lawless-m/claude-skills

Anthropic API rate limit handling - retry logic, backoff, throttling for batch workloads against Claude models

2026-07-276

auto-testing

lawless-m/claude-skills

Use when building an automated test → issue → fix loop with Claude Code and GitHub issues — overnight auto-fixing, regression loops, self-healing CI.

2026-07-276

blog-management

lawless-m/claude-skills

Use when creating, editing, publishing, or deleting posts on Cyril's Workshop blog or the steponnopets.net devblog.

2026-07-276

boofuzz

lawless-m/claude-skills

Use when writing or contributing a boofuzz network-protocol fuzzer in this repo — layout, formatting rules, and reading results.

2026-07-276

browserbridge

lawless-m/claude-skills

Use when a task needs real-time control of a connected browser via the Browser Bridge Broker — submit JS jobs over HTTP that browsers eval and return.

2026-07-276

character-lora-on-runpod

lawless-m/claude-skills

Use when training a character LoRA (Chroma/Flux or Pony/SDXL) on a RunPod GPU and wiring it into the ComfyUI + pony_web render stack.

2026-07-276

name	Qwen-Ollama
description	Use when a task needs local LLM inference via the on-box Ollama server (Qwen 2.5) instead of a cloud API.

Qwen via Ollama

Chosen model: qwen2.5:7b (4.7 GB). Other sizes exist if needed: 0.5b / 14b / 32b.
Endpoint: http://localhost:11434 (standard Ollama API).
Timeouts: 120 s default for analysis calls; longer (e.g. 300 s) for heavy generation.
Prefer local over cloud when: offline, data can't leave the machine, or bulk work where API cost matters.
VRAM contention with other GPU services: see the Vram-GPU-OOM-memory-management skill.

Reference implementation (Marvinous)

OllamaClient.rs in this folder — copy of the production async client (reqwest, timeout, error handling) from /home/matt/Marvinous/src/llm/client.rs.
Prompt building: /home/matt/Marvinous/src/llm/prompt.rs
System prompt (Marvin's personality): /etc/marvinous/system-prompt.txt
Config (model/endpoint settings): /etc/marvinous/marvinous.toml