Autonomous Agent Patterns
技能 活跃Design patterns for building autonomous coding agents. Covers tool integration, permission systems, browser automation, and human-in-the-loop workflows. Use when building AI agents, designing tool APIs, implementing permission systems, or creating autonomous coding assistants.
To provide developers with robust design patterns and implementations for creating sophisticated autonomous AI coding agents.
功能
- Agent loop and architecture patterns
- Tool design with schema and execution
- Permission and safety system patterns
- Browser automation and visual agent patterns
- Context management and checkpointing strategies
使用场景
- Building new autonomous AI agents
- Designing robust tool APIs for LLMs
- Implementing permission and security systems for agents
- Creating browser automation workflows for AI
非目标
- Providing a fully runnable agent out-of-the-box
- Specific LLM model integrations beyond conceptual examples
- Detailed deployment or scaling strategies
实践
- Agent Design
- Safety
- Tool Design
- LLM Integration
先决条件
- Python 3.8+
- Access to an LLM API (conceptual)
Trust
- warning:Issues AttentionIn the last 90 days, 17 issues were opened and 4 were closed, indicating a closure rate of approximately 23.5%, which is below the preferred 50% threshold.
安装
npx skills add davila7/claude-code-templates通过 npx 运行 Vercel skills CLI(skills.sh)— 需要本地安装 Node.js,以及至少一个兼容 skills 的智能体(Claude Code、Cursor、Codex 等)。前提是仓库遵循 agentskills.io 格式。
质量评分
类似扩展
Du Dum: Batch Then Act Pattern
96Separate expensive observation from cheap decision-making in autonomous agent loops using a two-clock architecture. A fast clock accumulates data into a digest file; a slow clock reads the digest and acts only when something is pending. Idle cycles cost nothing because the action clock returns immediately after reading an empty digest. Use when building autonomous agents that must observe continuously but can only afford to act occasionally, when API or LLM costs dominate and most cycles have nothing to do, when designing cron-based agent architectures with observation and action phases, or when an existing heartbeat loop is too expensive because it calls the LLM on every tick.
Cli Creator
100Build a composable CLI for Codex from API docs, an OpenAPI spec, existing curl examples, an SDK, a web app, an admin tool, or a local script. Use when the user wants Codex to create a command-line tool that can run from any repo, expose composable read/write commands, return stable JSON, manage auth, and pair with a companion skill.
Context Mode Ops
100使用并行子代理军队管理 context-mode GitHub 问题、PR、发布和营销。为每个任务编排 10-20 个动态代理。在分类问题、审查 PR、发布版本、撰写 LinkedIn 帖子、宣布发布、修复错误、合并贡献、验证 ENV 变量、测试适配器或同步分支时使用。
Fixflow
100使用严格的交付工作流执行编码任务：构建完整计划、分步实现、持续运行测试，并默认在每一步 (`per_step`) 后提交。当用户要求行为驱动交付或需求不明确时，支持显式提交策略覆盖 (`final_only`, `milestone`) 和可选的 BDD（给定/当/则）。
Kotlin Mcp Server Generator
100Generate a complete Kotlin MCP server project with proper structure, dependencies, and implementation using the official io.modelcontextprotocol:kotlin-sdk library.
Spec to Repo
100Use when the user says 'build me an app', 'create a project from this spec', 'scaffold a new repo', 'generate a starter', 'turn this idea into code', 'bootstrap a project', 'I have requirements and need a codebase', or provides a natural-language project specification and expects a complete, runnable repository. Stack-agnostic: Next.js, FastAPI, Rails, Go, Rust, Flutter, and more.