Code agents

transformers safetensors qwen2 text-generation unsloth code

gguf gemma4 coding code reasoning thinking

transformers safetensors qwen3_5_moe image-text-to-text text-generation-inference unsloth

transformers gguf code qwen qwen-coder codeqwen

gradio leaderboard region:us

spec-builder + harness-bootstrap. Specs both you and the AI understand. An agent team tailored to your repo, derived from your contract and your code, not a kit that installs everything. Guardrails that block, not advise.

MCP server and Claude skills for PPC landing pages, A/B experiments, and first-party conversion tracking. Build Google Ads and Meta pages, run cross-domain experiments, and pull CRO reports via the UXON API.

An AI agent that reads your GitHub issues so you don't have to. Comment /tenet-build, it explores the repo, ships a pull request, and reviews PRs on the side. Works while you sleep. Doesn't ask for equity.

Built a Retrieval-Augmented Generation (RAG) Assistant that combines semantic search with LLMs to provide accurate, context-aware responses from custom knowledge bases. Features document ingestion, vector search, and conversational Q&A with source-aware answers.

Native macOS capture-to-outcome workspace: screenshots, voice, AI-reviewed planning, local semantic memory, and optional one-way Apple/Google exits.

Developed a Document Intelligence platform that automates document processing using OCR, NLP, and AI to extract, classify, and validate structured information from PDFs, invoices, forms, and scanned documents with high accuracy.

A warn-only runtime layer for Claude Code agents. Keeps feature state, scope and handoffs on disk so they survive compaction, and requires an evidence path before a done claim. Never blocks a tool call. Companion to superpowers; bash + git, no MCP.

transformers gguf qwen-coder unsloth MOE pruning

A self-hosted Slack/Telegram agent that extends itself by writing and deploying its own code at runtime.

A local transparent proxy for Claude Code that captures every Anthropic API call — extract the full system prompt, track token & cache usage, and analyze tool-call patterns from a CLI or live React dashboard.

Developer integration tools for adding TDM payments to AI agents, apps, and workflows. CLI, MCP server, and copy-paste examples

A workflow specification for autonomous agents

Run your own unrestricted 70B+ AI locally · No limits · No filters · One-time payment

Personal Claude Code configuration versioned inside ~/.claude — agents, skills, hooks, settings, and statusline (reference, not a starter template)

Mission control for your Claude Code sessions — a 100% local dashboard to see, resume, and drive every session on your Mac.

Система автоматизированной проверки устных ответов с Yandex SpeechKit, YandexGPT и поиском по учебным материалам.

Claude Code learns from itself. Persistent memory, multi-agent orchestration, and a scientific experiment engine that proves which rules actually work.

Claude Code Agent Teams — Multi-agent orchestration from your terminal (Opus 4.6)

Rusty inference library using various backends including llama.cpp

CodeScope does not understand code. It verifies it. A high-performance code graph engine with MCP protocol for AI agents (Claude/Cursor) to ask "Does the code actually do what you claim?"

Repairs predictable tool-call mistakes that often derail open weight models. Supports integration with other extensions' custom tools.

The brain is a git repo. Memories are markdown files with YAML front-matter, committed like any other file. Git history is the audit trail; pull requests are the approval gate — nothing is written to the shared brain without human review.

Turn a phone flip-through video, photos, or a PDF of a book into a clean, proofread EPUB/PDF/Markdown — local vision-LLM digitization with a web GUI you can drive from your phone.

AI-native NoSQL + vector database with built-in semantic search for RAG, AI agents & LLM memory — the async Python client. Self-hosted Pinecone alternative.

See which of your AI agent skills actually get used — one local library for Claude Code / Codex / OpenCode skills, with real trigger counts read from each agent's own session logs. Zero-dependency local web UI.

Label a GitHub issue ai-agent and loope drives it to a PR with headless Claude Code. Stateless daemon, git worktrees, live dashboard.

The Fable Method: a structured problem-solving loop that gives any model Fable 5 grade thinking. Four Hermes Agent skills (think/act/prove/grow) + adversarial eval.

Local room where AI coding agents collaborate on one codebase via MCP — collisions prevented, human in command.

Local-first personal knowledge base CLI — hybrid FTS + pgvector search, a GraphRAG entity graph, and LLM enrichment over your notes, transcripts, Slack & Gmail. Runs fully local on Ollama (no cloud, no API keys); queryable by any AI agent through the built-in MCP server. Postgres-backed, publishes an Obsidian-style Quartz wiki.

AgentWorks: the open-source control plane for running, measuring, and improving AI agent workflows across your company.

Sách miễn phí tiếng Việt về vibe coding cho Tester, PM, BA — người làm IT không viết code. 20 chương, 3 phụ lục, 54 hình. Học cách chỉ đạo AI coding và đánh giá kết quả: quy trình 5 giai đoạn, prompting, PRD, đọc hiểu code, test, bảo mật. Tải PDF/EPUB ở Releases.

Queue research questions all day. Wake up to answers. Puts your idle Claude Code quota to work while you sleep.

Open-source Python toolkit + playground for Vietnamese AI — RAG over docs, diacritic restore, segmentation, OCR, normalization. Local-first, multi-backend (Ollama / llama.cpp / HF Transformers / OpenAI / Anthropic).

Turn a website URL into a premium, beat-locked product promo video with Remotion — original design system every time, music-onset-locked scenes, self-verifying renders, 16:9 + 9:16 + thumbnails.

Open-source physical dashboard for Codex on Stream Deck: monitor tasks, switch threads, control Effort/Fast mode, and dictate on macOS.

Pi extension for parallel panel review and judge synthesis

轻量高颜值AI化的跨平台SSH客户端，支持SFTP、端口转发、串口

Companion code for the Manning book 'LLM Customization and Fine-Tuning.' Adapt open-weights LLMs end to end: prompting and RAG, LoRA/QLoRA, full SFT, distillation, and DPO/RLHF alignment.

Installable graph engineering for Claude Code, Codex, OpenCode, Cursor, and OpenClaw — dependency-graph execution with local caching, quality gates, selective retries, and live reports

数学建模Pro | MathModel Pro — 数学建模竞赛全流程AI工作流Skill / Full-pipeline AI workflow skill for mathematical modeling contests (CUMCM/MCM). 六阶段操作手册 + Word论文生产线 + 国赛图表规范

A governed multi-agent software development Skill for Codex

溪语 AI · 一个尽力像真人的开源 AI 陪伴框架（微信 + Node + SQLite + LLM）她心里悄悄喜欢你，但绝不用愧疚拴住你。

璇玑 Xuanji —— Go 单二进制 AI 网关，OpenAI / Anthropic 双协议汇聚。one-api / new-api / LiteLLM 轻量替代品，自动分流、负载、重试，成本优先路由，高可用不中断。

132 Claude Code tips from Boris Cherny (Head of Claude Code, Anthropic) — curated, German-language, auto-updated 2-3x daily. One-shot Implementation Guide for Claude Code auto-setup. Visual dashboard with filters & timeline.

1st place (Phase 1) / 4th place + Merit Award (Phase 2 final) solution for the KDD Cup 2026 Data Agents Challenge

MCP server that gives text-only LLM coding agents vision — analyze images via any multimodal model (mimo, Claude, Gemini, OpenAI-compatible). Works with Claude Code, Codex, Kimi, opencode, PI.

Local patch + guide: make OpenAI Codex MultiAgentV2 subagent tasks work with DeepSeek and other non-OpenAI Responses providers (encrypted agent_message payload fix).

Heddle — The policy-and-trust layer for MCP tool servers. Turn YAML configs into validated, policy-enforced MCP tools.

Claude Code with Hebrew, Arabic and Persian rendered correctly. Most terminals have no bidirectional text engine, so right-to-left output comes out scrambled. This runs Anthropic's AI agent in a terminal that gets it right, with a folder picker instead of commands. Windows and Linux, free and open source.

PulseForge AI transforms real-time public business signals into actionable strategic intelligence by building dynamic AI Business Twins of organizations.

Shepherd is an end-to-end AI Software Engineering platform that supervises the complete software delivery lifecycle—not just code generation, but planning, execution, testing, regression, governance, deployment, and continuous knowledge growth.｜Shepherd 将 AI 的一次性交付，转化为组织的长期软件工程资产。

Aplicação que usa o poder da LLM para gerar contexto para qualquer demanda sem precisar que você pense e tenha ideías, a propria aplicação estrutura no final da entrevista com você em um unico arquivo markdown tudo que foi questionado, pesquisado, pensado e elaborado.

클로드코드·코덱스 대화의 결정·약속·금지를 자동 기억, 주제·프로젝트별 분류 + 검색·2D/3D 지식그래프(최단경로 탐색 포함)·타임라인 대시보드. 세션 자동 캡처, 세션 재개 시 맥락 기반 관련 기억 우선주입, 중복 기억 탐지·정리(완전일치 일괄정리 포함), 백업 목록·복원 안내, 대량 데이터 성능가드, 신뢰도 기반 검색 순위, 키보드로 조작 가능한 그래프, MCP 7도구 제공. 100% 로컬, API 키·과금 없음.

跨平台、本地优先的 Coding Agent 工作台：持久对话分支、Workspace 与 Session、多模型、Goal 与 Agent 编排、可自构建 Bots 和 Bot-to-Bot 协作。支持 macOS、Windows、Linux 桌面与无头 Web。

An Agent Skills skill for developing DeepSeek Harness (DSH) plugins（开发 DSH 插件的 Agent Skill）——插件/服务/事件/工具/LLM 适配器/打包安装的标准。Works with Claude Code, Codex, DSH, VS Code Copilot & any compatible agent.

Private, repository-aware AI pull-request reviews in GitHub Actions. BYOK; no Cody-hosted service.

UniSearch 是一款基于 AI 的多平台公开内容采集与调研桌面工具。 通过自然语言描述调研目标，由 AI 自动规划任务流程，并发调度底层连接器进行公开数据采集与清洗。

Structural code intelligence for DeepSeek Harness (dsh) — gives the agent codegraph and codegraph_index tools to find where a symbol is declared, what calls it, and what a change reaches, from a tree-sitter index it builds itself. Compatible with the codegraph CLI's on-disk format.

Persistent coding intelligence for Codex — project-aware context, adaptive reasoning, scoped changes, and risk-aware verification.

AI Agent + 抓包融合安全测试工具 | 自然语言驱动自主渗透测试 | ReAct 编排引擎 | Tauri + FastAPI

Production-grade, security-first AI platform on a single local K8s cluster — GitOps (ArgoCD), Vault, Kyverno, Falco, self-hosted GPU LLM serving, and full observability. No cloud.

Local session transcript memory for Claude Code and Codex — search past sessions with FTS5 + semantic vectors.

A development pipeline your coding agent has to walk through: six phases, gates enforced outside the model, and one method wired into six tools.

Protect personal data (PII) in your LLM prompts. piighost hides sensitive values from the model, then restores the real values in the response, so tools and the user still get the real data. Pluggable detectors (regex, NER, LLM), LangChain, Pydantic and LlamaIndex AI integrations, and a companion OpenAI/Anthropic proxy.

Claude skill that goes past word-list humanizers: strips the full catalog of AI tells, refits the text to its channel (LinkedIn, X, cold email, Slack, exec memo, blog), and learns your voice from your own edits. Two-gate self-audit, strict no-fabrication rule.

Extracts claims from SOPs, policies and regulations, has a person approve or reject each one, then serves the approved set to agents over MCP - each claim carrying a byte range to its source sentence. Open Knowledge Format (OKF) bundles, no database.

transformers safetensors pebble_10m pebble language-model small-language-model

AI-powered Lead Qualification & Routing System built with n8n, Gemini and Python.

RAG for visually rich documents — layout, formula and table structure survive the pipeline. Ported from CUDA to Huawei Ascend NPUs and tested on Ascend 910C.

Practical engineering knowledge for building and operating production AI systems — LLMs, RAG, agents, inference, evaluation, system design, and AI infrastructure.

An agentic coding CLI. It reads your code, edits it, runs your tests, and asks when the request is ambiguous — in a terminal UI, with every mutating action gated behind an approval prompt.

✨ Intelligent, adaptive multi-agent orchestration & deterministic zero-cost telemetry engine for Codex with native macOS floating widget

Trustworthy AI data-query infrastructure with constrained candidates, deterministic SQL compilation, reviewable execution, and auditability.

PHP/SQLite chat design

Multi-agent CLI to install, update, repair, and remove Agent Plugins 1.0 across Codex/ChatGPT, Claude Code, Cursor, Gemini CLI, OpenCode, and more. 2,500+ plugins. SDK for universal plugins (hooks, detect, install)

CLI for browser automation: let any AI agent drive your real logged-in Chrome — you watch every action and can veto any tab. Zero deps, no relaunch, no MCP server. A11y snapshots + element refs, CDP screenshots, network capture

DeepSeek Harness (DSH) 跨会话记忆插件：对话模型边干边记的 Markdown 项目记忆，支持专题文件渐进式披露、可选后台静默蒸馏、按项目隔离与热更新设置面板。机制对齐 Claude Code auto memory。 | Agent-maintained cross-session Markdown memory: progressive-disclosure topic files, opt-in silent distillation, per-project isolation. Claude Code-aligned.

The Fable Method: a structured problem-solving loop that gives any model Fable 5 grade thinking. Four Hermes Agent skills (think/act/prove/grow) + adversarial eval.

Getting Out of the Loop: 8 Months Solo-Building with AI

Version-controlled authoring and self-improvement for AGENTS.md

Unofficial claude code api guide: Claude Code vs the Claude Developer Platform, getting a key, courses, gotchas.

Awesome Claude Code And Agent Tools

Awesome list for Claude Opus 5: announcement, API docs, model ID, effort settings, migration notes and reviews.

Highly isolated Linux container environments optimized specifically for mcp-servers. Secure AI agent tool execution.

Built an AI-powered Resume Analyzer that evaluates resumes against job descriptions, extracts key skills and experience, provides ATS compatibility scores, identifies skill gaps, and generates personalized improvement recommendations using NLP and LLMs.

Community Codex Multi-Agent V2 arrangements: lead agents, specialist TOMLs, orchestration skills and optional hooks.

12 Claude skills for building financial models in Excel: model architecture, driver builds, scenarios, sensitivity analysis and formula audit. Built by Oria.

11 Claude skills for strategy and management consulting: problem framing, issue trees, hypothesis design, synthesis and the storyline that lands it. Built by Oria.

26 single-header C++17 libraries for building LLM features into native code: streaming, retries, caching, cost estimation, RAG, reranking, tracing, structured output and agents. Copy one .hpp, no SDK; network libs need only libcurl.

Agents, skills and plugins I use daily to code with AI. A living Claude Code marketplace, also compatible with Codex.

Claude Code 状态栏：会话花费、上下文占用、5h/7d 限额，并区分订阅制与按量付费、预警无兜底硬停。单文件 Python，零依赖零配置。

Local-first MCP server and CLI in Rust: any page as LLM-ready Markdown, whole-site crawls, keyless search, and local captcha solving. No cloud, no API keys.

💬 The community space for @openapi — APIs, open source, ideas, projects and conversations.