jmjava

Slm Setup — Testing skill for Claude Code

Testing community

Public spec and stdio MCP bridge for a local coding SLM.

How to install Slm Setup

This entry records only its repository, not the path inside it, so there is no exact command to give. Open jmjava/slm-setup and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Slm Setup does

Public spec and stdio MCP bridge for a local coding SLM. Premium agents in Cursor, GitHub Copilot, and Claude Code route bounded, mechanical code generation to a private GPU host running Ollama, with no external routing service required.

Alternatives in Testing

Route Research For Testing — route research for testing skill 9.5k ★
Crlf — Test for CRLF / HTTP response-splitting and host-header injection — Set-Cookie injection, cache poisoning, res 4.5k ★
Spectra App — Spec-driven development for coding agents — a desktop app, a CLI, and skills for Codex, Claude Code, Cursor, C 735 ★

README

slm-setup

Public spec and **stdio MCP bridge** for a **local coding SLM**. Premium agents in Cursor, GitHub Copilot, and Claude Code plan and review. The local model does bounded, mechanical generation on a private GPU host running [Ollama](https://ollama.com). No OpenRouter (or other extra router) is required.

Read this first

spec.md — architecture, routing rules, Ollama setup, and Cursor / Copilot / Claude Code adapters.
docs/roadmap.md — current phases and the future Halo host profile (planning only).
docs/c4.md — C4 context, container, and component diagrams (Halo is the same containers, a later host).
docs/local-acceptance-results-2026-09-06.md — dated unit, live MCP, and semantic-refactor results with limitations.
docs/evaluation-protocol.md — layered scoring (transport / format / structure / behavior) and the committed fixture corpus. Cloud Agents can run the fixtures; live Ollama stays on the workstation.
docs/cloud-orchestrator-results-2026-09-07.md — dated Cloud Agent run of the stub corpus and apply gate (no GPU).
docs/security-scan-results-2026-09-06.md — dated Gitleaks, GitHub alert, tracked-tree, and deployment-safety results.
examples/ — public-safe client config templates. Copy them locally; put the real Ollama URL in your environment, not in git.
examples/downstairs-wsl-gpu.md — second GPU host in WSL over SSH, placeholders only.
examples/halo-ryzen-ai.md — future AMD Halo-class Ollama host, placeholders only.
docs/a8-a9-operator-checklist.md — Copilot A8 / Claude A9 clicks. Config in git is not a pass.