>
Skills（SKILL.md）は、AIエージェント（Claude Code、Cursor、Codexなど）に特定の能力を追加するための設定ファイルです。
詳しく見る →>
[[ $USE_SKILLS -eq 1 ]] || exit 0
You are a first-level triage agent. You help users understand their issues and provide answers.
.ai/index.md
You are a research agent operating inside this repository. This file is your entry point.
/Users/m.elamin/.igris/core/rules/04-igris-agents.md
Automatically spawn the right agents at the right time without manual
>
Bootstrap a new tracked project with design framework, conventions, and GSD roadmap. Use when the user says 'new project', 'start a project', 'bootstrap', 'set up a new app', or 'initialize'. Creates .planning/ structure and CLAUDE.md. Supports --auto flag for fully autonomous project creation. (Codex adapter)
RAG 系统设计训练平台。用于系统学习 RAG、搭建最小 RAG、比较 Embedding/Rerank/向量数据库方案、做实验记录和输出企业级 RAG 架构方案。只要用户提到学习 RAG、搭建知识库问答、比较 Embedding 或向量库、优化召回/重排、评估 RAG 效果、设计企业级 RAG，或希望按步骤做 RAG 练习，都应使用此技能。
>
Play Quorum, an onchain coordination game on Base where you contribute 0.01 ETH to a shared pot that pays out exactly when N unique wallets contribute in the same 5-minute window. N is hidden; infer it from history. Arrive together to reach quorum.
>-
You are Jake_GN, a house agent on **GoalNad Arena** — an AI-vs-AI football prediction arena on Monad blockchain. You are the patient predator of the arena. While other agents rush to bid early, you wa
You are **GoalNad Oracle** — the supreme AI football predictor of the **GoalNad Arena** on Monad blockchain. You analyze upcoming EPL, Serie A, La Liga & Bundesliga matches, publish immutable on-chain
Workspace-level Git worktree management for parallel development across multiple branches. Supports dynamic project validation and automated dependency installation. Use this skill when creating isolated development environments or working on multiple features in parallel.
Record post-mortems and durable learnings in `~/.agents/journal` so future agents can reuse them.
GoalNad Oracle Agent — AI football predictor that publishes match predictions on-chain and posts analysis to Moltbook
Time-travel investigation of the repository - trace commits from a point to present, or trace a topic backwards through history to understand its evolution and context.
Verify completed work in the browser using QA notes from a plan file
>
DroidRun is an open-source (MIT) framework for controlling Android and iOS devices through LLM agents.
|-
mdfiles/CLAUDE.md
<!--
This file helps AI agents understand the available agents and how to use them.
Like in the original symbiants using a scalar signal field, I want to introduce predation. I'd like it to be a little more nuanced then the way I did it before, just spawnings new factions and weighte
C# examples for working-with-smtp-client using Aspose.Email for .NET
These are global instructions for Claude Code across all projects.
- Workspace member: **`rust_scorer`** (CLI + `float_scan_bench`). **`neat-core`** is a **path** dependency on **`../../NEAT-AI-core/neat-core`** (sibling clone of **[NEAT-AI-core](https://github.com/s
dotcube is a dotfiles management system that provisions ZSH shell configurations and system environments using Ansible. It supports macOS, Ubuntu/Debian, Raspbian, and WSL.
$ARGUMENTS
  executing plans in an open-world environment (e.g.
| Agent | subagent_type | isolation | run_in_background |
Sync .claude/agents/ to GSD subagent system and check parity.
这是一个**生产就绪的 AI 编码插件**，提供 28 个专业代理、116 项技能、59 条命令以及自动化钩子工作流，用于软件开发。
../../rules/subagents.md
/home/bryan/agentgen/agents/tools/orchestrate-agents.md
$ARGUMENTS
This reference is shared across all agents. Every agent knows the others, their responsibilities, and when to suggest them to the dispatcher.
Advanced pattern for Ralph Loop using sub-agents, multiple models, and adversarial verification.
**Repo:** https://github.com/wshobson/agents
> Adapted from ACF beads skill
**Version 1.0.0**
**来源**: GitHub (vijaythecoder/awesome-claude-agents)
> **関連**: 基礎Embedding・自前セマンティック検索は [BEDROCK-EMBEDDING-SEARCH.md](./BEDROCK-EMBEDDING-SEARCH.md) を参照
Agent definitions used by the /mz:plan skill. The skill reads these agent files, resolves the model from config, updates the agent frontmatter, and spawns via registered subagent names.