| name | openai-docs |
| description | Use when the user asks how to build with OpenAI products or APIs and needs up-to-date official documentation with citations (for example: Codex, Responses API, Chat Completions, Apps SDK, Agents SDK, Realtime, model capabilities or limits); prioritize OpenAI docs MCP tools and restrict any fallback browsing to official OpenAI domains. |
| knowledge_graph_profile | references/task-profile.json |
OpenAI Docs
Provide authoritative, current guidance from OpenAI developer docs using the developers.openai.com MCP server. Always prioritize the developer docs MCP tools over web.run for OpenAI-related questions. Only if the MCP server is installed and returns no meaningful results should you fall back to web search.
Quick start
- Use
mcp__openaiDeveloperDocs__search_openai_docs to find the most relevant doc pages.
- Use
mcp__openaiDeveloperDocs__fetch_openai_doc to pull exact sections and quote/paraphrase accurately.
- Use
mcp__openaiDeveloperDocs__list_openai_docs only when you need to browse or discover pages without a clear query.
OpenAI product snapshots
- Apps SDK: Build ChatGPT apps by providing a web component UI and an MCP server that exposes your app's tools to ChatGPT.
- Responses API: A unified endpoint designed for stateful, multimodal, tool-using interactions in agentic workflows.
- Chat Completions API: Generate a model response from a list of messages comprising a conversation.
- Codex: OpenAI's coding agent for software development that can write, understand, review, and debug code.
- gpt-oss: Open-weight OpenAI reasoning models (gpt-oss-120b and gpt-oss-20b) released under the Apache 2.0 license.
- Realtime API: Build low-latency, multimodal experiences including natural speech-to-speech conversations.
- Agents SDK: A toolkit for building agentic apps where a model can use tools and context, hand off to other agents, stream partial results, and keep a full trace.
If MCP server is missing
If MCP tools fail or no OpenAI docs resources are available:
- Run the install command yourself:
codex mcp add openaiDeveloperDocs --url https://developers.openai.com/mcp
- If it fails due to permissions/sandboxing, immediately retry the same command with escalated permissions and include a 1-sentence justification for approval. Do not ask the user to run it yet.
- Only if the escalated attempt fails, ask the user to run the install command.
- Ask the user to restart Codex.
- Re-run the doc search/fetch after restart.
Philosophy
- Prefer root-cause understanding over quick symptom patches.
- Keep guidance evidence-based, explicit, and reproducible.
- Optimize for decisions that reduce rework and operational risk.
Workflow
- Clarify the product scope (Codex, OpenAI API, or ChatGPT Apps SDK) and the task.
- Search docs with a precise query.
- Fetch the best page and the specific section needed (use
anchor when possible).
- Answer with concise guidance and cite the doc source.
- Provide code snippets only when the docs support them.
Quality rules
- Treat OpenAI docs as the source of truth; avoid speculation.
- Keep quotes short and within policy limits; prefer paraphrase with citations.
- If multiple pages differ, call out the difference and cite both.
- If docs do not cover the user’s need, say so and offer next steps.
Tooling notes
- Always use MCP doc tools before any web search for OpenAI-related questions.
- If the MCP server is installed but returns no meaningful results, then use web search as a fallback.
- When falling back to web search, restrict to official OpenAI domains (developers.openai.com, platform.openai.com) and cite sources.
Anti-patterns
- Skipping investigation and jumping directly to fixes.
- Making claims without evidence, logs, or reproducible steps.
- Mixing unrelated workstreams in a single execution path.
Constraints / Safety
- Redact secrets, tokens, credentials, and PII by default; never echo raw environment values.
- Prefer safe defaults and avoid irreversible changes without explicit confirmation.
Inputs
- User task context and target environment.
- Relevant constraints, permissions, and preferences required to execute safely.
Outputs
- A concrete next-step response with explicit, reproducible actions.
- A short verification checklist and caveats for the user.
Validation
- Fail fast: stop at the first failed check and do not continue.
- Re-run the required checks before proceeding to the next step.
- Report any failed check and requested follow-up actions clearly.
When to use
- Use this skill when the request matches the skill's intent and scope.
- Do not use it when a different domain or higher-privilege workflow is required.
Constraints / Safety
- Redact secrets, tokens, credentials, and PII by default; never echo raw environment values.
- Prefer safe defaults and avoid irreversible changes without explicit confirmation.
Philosophy and tradeoffs
- Use this skill when consistent decision-making matters more than one-off execution because project context should drive the approach.
- Principle and mindset: prioritize tradeoffs and constraints over rigid checklists; understand why each step exists.
- Ask this to keep outcomes robust: Why is this the right default, and what could change this outcome?
- How do we adapt if constraints shift?
- What evidence is needed before choosing one path over another?
Anti-patterns and caveats
- Avoid applying this playbook generically without checking repository-specific context.
- NEVER skip required validation gates when behavior changes.
- DO NOT use this skill as a rigid replacement for engineering judgment.
- DON'T ignore warnings or assume one pattern fits all repos.
- Common pitfall: treating anti-patterns as optional.
- Incorrect assumptions here can lead to fragile guidance.
- Warning: wrong sequencing can create avoidable regressions.
Variation and adaptation
- Vary the workflow by team size, risk, and deployment target.
- Use different strategies for small, medium, and large changes.
- Adapt recommendations to the specific environment and avoid repetitive templates.
- Avoid generic or cookie-cutter responses; craft context-specific alternatives.
- Keep outputs diverse and not repetitive.
- Converge on a custom path only after evidence review.
- Different constraints should produce different, non-generic recommendations.
Empowering execution style
- Be capable of exploring multiple options and enabling the team to make safe decisions.
- Unlock confidence by explaining options and tradeoffs clearly.
- Feel free to be creative while staying rigorous and precise.
- Push boundaries with practical alternatives when simple recipes fail.
- Enable outcomes-oriented problem solving.
