however-yir/codex-skill-ai-agent-workflow
Codex skill for repeatable AI agent workflows with prompts, tools, MCP integrations, and evaluation loops.
🔥 A Codex skill for building repeatable AI agent workflows with prompts, tools, and MCP integrations. 🚀 Designed for workflow decomposition, tool grounding, evaluation loops, and reusable execution specs. ⭐ Helps teams move from one-off prompts to production-minded agent systems.
[](./SKILL.md) [](./LICENSE) [](https://github.com/however-yir/codex-skill-ai-agent-workflow) [](https://github.com/however-yir/codex-skill-ai-agent-workflow/commits/main)
> Repeatable AI workflows with prompts, tools, MCP integrations, and evaluation loops.
中文简介：这是一个面向 AI 工作流设计的 Codex skill，适合把一次性提示词任务沉淀成可复用的 prompts、tools、MCP 集成和评测闭环。
A Codex skill for turning one-off AI ideas into repeatable workflows with prompts, tools, MCP integrations, checks, and evaluation loops.
Use this when the main stack direction is already mostly known. If the bigger question is still local model choice, deployment path, or LM Studio versus Ollama versus MLX, local-ai-systems-studio is the better first stop. If the main request is to create, rewrite, benchmark, or optimize a skill itself, skill-creator is the better fit.
This skill helps design AI workflows that can be reused, inspected, and improved over time. It is built for prompt pipelines, tool-using agents, workflow decomposition, and deciding when something should become a prompt stack, script, or MCP server once the stack direction is already clear.
Loading reviews...