Communitygithub.com

dayfinggg/openai-codex-agent-skills

Production engineering and design skills for OpenAI Codex.

openai-codex-agent-skills 是什么?

openai-codex-agent-skills is a Codex agent skill that production engineering and design skills for OpenAI Codex.

兼容平台~Claude CodeCodex CLI~Cursor
npx skills add dayfinggg/openai-codex-agent-skills

在你喜欢的 AI 中提问

打开一个已预加载此 Agent Skill 的新对话。

文档

Use only when proceeding would commit to a high-impact choice that local evidence cannot settle. High-impact choices affect public behavior, data meaning, security or privacy, compatibility, resource limits, architecture, destructive actions, or external writes.

Do not use for routine fixes, reversible details, reviews, explanations, or decisions already fixed by the request or repository guidance. Use align-requirements for ordinary ambiguity.

Inspect the repository, tests, configuration, history, and authoritative documentation before asking about facts. Form one concrete hypothesis about the intended outcome. Ask the user to confirm or correct that hypothesis with one focused question. When the choice is bounded, give two or three mutually exclusive options and recommend the first only when evidence supports it.

Wait for the answer before asking a dependent question. Treat vague or conflicting answers as unresolved. Do not turn the interview into a broad questionnaire, explore hypothetical future branches, or edit files while a required decision is open.

Stop when the material decision is settled and return the confirmed decision plus any explicit assumption needed by the normal implementation workflow.

Read references/examples.md when the trigger boundary or first question is unclear.

相关技能