Skill：agent skills
Shared skills for AI agents (Claude Code, Codex, Gemini)
适用场景
教研备课
适合角色
介绍
以下内容整理自开源仓库 `grahama1970/agent-skills` 的 README / SKILL.md 摘要： # Agent Skills  **Bring me the problem before it has a specification.** I'm Graham Anderson. This is the working laboratory behind my agent engineering practice: reusable tools, research experiments, bounded workers, and the lessons I would rather not learn twice. You bring the situation, the constraints, and room to investigate. I work out what actually needs solving, test the uncertain parts, and build. Defining the problem is part of the work—not homework you need to finish before we talk. [Explore the work](https://grahama.co) · [Start here](#five-places-to-start) · [Browse the skills](skills/) · [About me](RESUME.md) · [Get in touch](#bring-me-the-project-you-shelved) > **Public source, partly private runtime.** This is a working research toolbox, > not a turnkey SDK. Some paths need configured memory services, model providers, > browser bindings, credentials, or media storage. You can inspect the public > code and contracts without having my runtime; you cannot assume every workflow > will run in a fresh clone. | If your problem involves… | Start with | |---|---| | Reliable agent workflows | [Tau](https://github.com/grahama1970/tau) | | Proving whether automation works | [agentic-evals](skills/agentic-evals/README.md) | | Adversarial validation | [Battle](skills/battle/README.md) | | Organizational memory and provenance | [Memory](https://github.com/grahama1970/memory-public) | | Voice, personas, or multimodal systems | [Persona Dream](skills/persona-dream/README.md) | | Something unclear or cross-cutting | [Contact Graham](#bring-me-the-project-you-shelved) | Typical engagements start with one of four shapes: - investigate an underspecified technical problem; - prototype and test the risky part; - repair or harden an existing agent workflow; or - build an evidence-backed implementation and handoff. ## Five places to start Pick the question that interests you. These are routes into the work, not a ranking of finished products. ### How do you let agents act without accepting their word for what happened? **[Tau](https://github.com/grahama1970/tau)** explores that problem through bounded execution, typed handoffs, evidence checks, and explicit stop conditions. The interesting part is not drawing a graph of agents. It is deciding what must be true before the next action is allowed. Start with the canonical repository and its current-status link. The [local Tau skill](skills/tau/README.md) is the agent-facing integration, not a second implementation of the whole system. Useful when an internal agent must act without bypassing review or losing a trace of what happened. ### What does a passing evaluation actually establish? **[agentic-evals](skills/agentic-evals/READM --- **说明**：内容由公开仓库整理，使用前请自行核对是否适合你的场景。
安装方式
1. 打开仓库：https://github.com/grahama1970/agent-skills 2. 查找 SKILL.md 或 README 中的安装章节 3. 克隆或下载所需 Skill 子目录到本地 skills 路径（如 `.cursor/skills/` 或 `.claude/skills/`） 4. 重启 / 重载 Agent 客户端 5. 用一条真实任务验证 Skill 是否被正确加载
适用平台
- Claude
- Codex