openai-api-troubleshooting — Codex skill: what it does and install
Diagnoses failed OpenAI API requests in Codex and pinpoints the immediate next action. Covers network blocks, invalid keys, depleted quota, rate limits, and model access issues.
npx skills add openai/plugins --skill openai-api-troubleshooting -a codexWhat openai-api-troubleshooting does
This skill handles runtime failures that occur after sending a request to the OpenAI API. It routes problems into five clear groups: sandboxed network blockers, authentication errors, credit depletion, rate limits, and access restrictions. Transport failures come first. When a sandboxed run hits DNS errors or connection resets, the skill tests the network route by retrying the call with sandbox_permissions=require_escalated.
The routing logic separates tricky edge cases. Both rate limits and an empty balance can produce HTTP 429, but the skill treats them differently. For rate_limit_exceeded, it suggests pacing or exponential backoff. For insufficient_quota, it directs you to organization billing and points to gpt-5.4-mini as an initial testing option. It leaves key provisioning to openai-platform-api-key and documentation lookups to openai-docs.
Explanations stay brief and actionable. Install this skill if you run automated tasks through Codex and want immediate clarity on API failures without digging through logs.
What it helps with
- diagnose runtime failures from OpenAI API requests
- distinguish insufficient_quota billing stops from rate_limit_exceeded pauses
- rerun sandboxed calls that hit DNS or connection timeouts with escalated permissions
- identify 403 model or organization access permissions
Example prompts for the agent
- My OpenAI API call failed with a 429 error, check if this is a rate limit or a billing issue
- The script cannot reach api.openai.com from the sandbox, help me fix the connection
- I received a 403 model_not_found error, tell me what permissions or project settings are missing
How to install openai-api-troubleshooting
1Skills CLI (Vercel)
npx skills add openai/plugins --skill openai-api-troubleshooting -a codex2Manually: put the files in place
.agents/skills/openai-api-troubleshooting/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add openai/plugins --skill openai-api-troubleshooting -a claude-code2Manually: put the files in place
~/.claude/skills/openai-api-troubleshooting/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/openai-api-troubleshooting/
Manually: put the files in place
.cursor/skills/openai-api-troubleshooting/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install openai/plugins openai-api-troubleshooting2Manually: put the files in place
.github/skills/openai-api-troubleshooting/SKILL.mdPlatform CLI
gemini skills install https://github.com/openai/plugins --path plugins/openai-developers/skills/openai-api-troubleshootingManually: put the files in place
.devin/skills/openai-api-troubleshooting/SKILL.mdThe legacy .windsurf/skills/ path still works
Claude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
From the same set201
The whole set →Similar
Questions about openai-api-troubleshooting
How do I install openai-api-troubleshooting in Codex?
Run npx skills add openai/plugins --skill openai-api-troubleshooting in your terminal.
What does openai-api-troubleshooting do?
It categorizes failed OpenAI API calls into network issues, authentication errors, quota exhaustion, rate limits, or project permission mismatches.
Does this skill create or rotate API keys?
No, creating and configuring API keys is delegated to the openai-platform-api-key skill.