ColossusForgeStudios Claim this profile

Single-file, dependency-free live dashboard for local llama.cpp/vLLM serving — GPU, throughput, KV/ctx, model library. Stdlib Python + one local-first HTML page.

Claude Code plugin that halts the agent at an unverified build - a Toyota andon cord for AI coding loops.

Translate Anthropic /v1/messages to any local OpenAI-compatible server. Point any Anthropic SDK client at a self-hosted model.

A solo dev's autonomous AI game studio — a UE5 game, an agent fleet, and a local model that trains on its own work. Curated, receipts-first.

Autonomous AI development loop for Claude Code with intelligent exit detection

ComfyUI-42lux-Hildegard-Refiner Tile-based refinement nodes for ComfyUI, built around the Hildegard-Refiner reference-latent scheme. Designed for FLUX.2 Klein.

Offline single-file parts/cost picker for GPU & AI rigs — live cost math, source-link per part, paste-a-URL→grab-the-price. Includes 1×/4×/8× RTX 3090 and 4× 5090 reference builds.

Rust LLM serving traffic tap + live SSE dashboard for llama.cpp/vLLM — the tap-based sibling of llm-serve-dashboard

Self-hosted AI workspace.

Beautiful, AI-native markdown IDE and LLM wiki

Extracted system prompts from Anthropic - Claude Fable 5, Opus 4.8, Claude Code, Claude Design. OpenAI - ChatGPT 5.5 Thinking, GPT 5.5 Instant, Codex. Google - Gemini 3.5 Flash, 3.1 Pro, Antigravity. xAI - Grok, Cursor, Copilot, VS Code, Perplexity, and more. Updated regularly.

AI agent skills published by NVIDIA

An open-source AI coding agent that lives in your terminal.

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Community recipes for serving LLMs on RTX 3090/4090/5090 CUDA gpus. Multi-engine (vLLM, llama.cpp, ik_llama) and model-agnostic. Currently shipping Qwen3.6-27B Qwen3.6 35B Gemma 4 26B Gemma 4 31B configs for 1× and 2× cards.

GLM-5: From Vibe Coding to Agentic Engineering

A private, self-hosted browser start page (Rangers, Mets, markets, AP/Reuters) that runs only on your machine and phones home to nobody.

llama.cpp fork with additional SOTA quants and improved performance

from vibe coding to agentic engineering - practice makes claude perfect

Reference architecture + consultation playbook for adopting AI agents safely: T1-T4 trust tiers, circuit-breaker autonomy, tiered deployment.

Versioned canonical artifact store: Postgres/pgvector + object store + librarian CLI + FastAPI + inotify wiki indexer, append-only audit, content-hash dedup.

Corpus -> QLoRA -> GGUF -> live /lora hot-load: a 4-stage LoRA adapter lifecycle for local llama.cpp servers, no restart.

A whole-transcript skeptic over a coding agent's session - catches later-contradicts-earlier.

Single-file tool-call reliability bench for OpenAI-compatible endpoints - finds the silent >5-tool drop cliff.

Four persona-divergent reviewers over one endpoint, with a Jaccard decorrelation metric and a labeled-bug benchmark.

MCP server giving AI assistants full access to Unreal Engine 5.7 projects — 65+ tools, C++ editor plugin, visual review loop

337 Claude Code skills & agent skills & plugins (30+ Agents, 70+ custom commands, 330+ skills, customizable references, scripts)for Claude Code, Codex, Gemini CLI, Cursor, and 8 more coding agents — engineering, marketing, product, compliance, C-level advisory, research, business operations, commercial & finance, and your daily productivity skills.

Claude Code is an agentic coding tool that lives in your terminal, understands your codebase, and helps you code faster by executing routine tasks, explaining complex code, and handling git workflows - all through natural language commands.

A single CLAUDE.md file to improve Claude Code behavior, derived from Andrej Karpathy's observations on LLM coding pitfalls.

A lightweight, powerful framework for multi-agent workflows

《动手学大模型Dive into LLMs》系列编程实践教程

Knowledge Engine for AI Agent Memory in 6 lines of code

An open source template for building cloud agents.

Self-evolving agent: grows skill tree from 3.3K-line seed, achieving full system control with 6x less token consumption

A Claude Code plugin that automatically captures everything Claude does during your coding sessions, compresses it with AI (using Claude's agent-sdk), and injects relevant context back into future sessions.

Local-first identity, memory, and secrets for AI agents. Portable state across models and harnesses.

A repository for data set creation code using the ZTRAX data set.

No description.