Search

A/B test AI agent skills, prompts, and tool definitions with LLM-as-a-Judge

Run Claude Code, Codex, Antigravity, Cursor Agent and OpenCode as one runtime — persistent sessions, multi-agent councils, an OpenAI-compatible endpoint, an MCP server, and an ACP agent any editor can drive.

Persistent memory for Claude Code. Survives compaction and session boundaries: anti-patch reconcile-on-write with LLM-judged de-duplication, a forced PROGRESS.md handoff, a live PLAN.md anchor with an enforced directive ledger, FTS5 search, MCP tools and a GUI. Pure stdlib Python, zero runtime dependencies.

A practical coding agent for terminal workflows.

交互式教程：20 节课，从「模型只会说话」讲到写出能装进真实 DeepSeek Harness 的插件。含可跑 demo 与在线版。

Claude Code plugin: universal radial-tree exploration engine. One tree skill + swappable presets (brainstorm / attack / design / code-audit) for divergent ideation, adversarial critique, and design-space exploration. 12 framings × hard-ban-on-incomplete-leaves × stable convergence.

笨办法学 Agent：从 60 行代码到一个完整的 harness。Learn Agent the Hard Way — 中文电子书，全书 32 个练习已完结。

Autonomous AI Agent that detects and fixes deprecated API calls in Python, JS, TS via LLM-powered Pull Requests

Boxdawn — AI agent observability. Detect waste in multi-agent traces.

Super Agent Web SDK — API + 嵌入式聊天组件

Server-side Portal helpers: define agent tools and serve them from one route, derive the browser-safe widget key, sign user identities.

Project-local Codex multi-agent workflow for Tantu Wuyou frontend projects.

Always-on dependency security monitoring for DeepSeek Harness (DSH) plugins: find exact installed or candidate transitive vulnerable paths and breaking updates, then route evidence to a DSH Agent.

Import pi / opencode sessions, chat history, and agents into DeepSeek Harness (dsh): slash commands, session-start migration prompt, and a one-click Sync button in the composer.

Long-term story memory for AI agents — novels, Galgame, interactive fiction. Store every word, recall plot detail on demand, fixed context cost.

Scan and sanitize attacker-writable on-chain text before it reaches an LLM. Zero-dependency prompt-injection guard for web3 AI agents — library, and a drop-in proxy.

Agente de QA de Orquestra -- corre en una Mac, toma corridas de la cola y reporta resultados vía la API (nunca toca Firestore directo)

AI model: mradermacher/Muse-Glimmer-30B-Tachibana-Agent-i1-GGUF

AI model: mradermacher/AgentMercury-Qwen3.5-4B-i1-GGUF

AI model: mradermacher/Muse-Glimmer-30B-Tachibana-Agent-GGUF