Skill llm openai
Skill joeylu/runtime-free-llm-integration-skills/LLM/skill-llm-openai
Runtime-free skills for coding agents to build fail-fast LLM provider integrations across OpenAI, Aliyun Bailian / DashScope, DeepSeek, and Gemini.
npx -y skills add joeylu/runtime-free-llm-integration-skills --skill skill-llm-openaiAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Runtime contract for direct OpenAI API integration. Use it to resolve exact OpenAI model IDs, Responses or Chat Completions URLs, request fields, reasoning, tools, structured output, image input, Image API calls, streaming, response mapping, pricing references, and logging. Model choice remains with the user or host project.
SKILL.md
4.2 KB, as published. Nobody here has run it
Skill LLM OpenAI
Purpose
Define how an agent integrates a model already named by the user or host project. This skill does not recommend, approve, rank, or automatically replace models during ordinary implementation.
Read Order
Read shared contracts first:
../_shared/route-key-schema.md../_shared/model-catalog-schema.md../_shared/pricing-matrix-schema.md../_shared/capability-matrix-schema.md../_shared/connection-profile-schema.md../_shared/request-url-matrix-schema.md../_shared/request-envelope.md../_shared/response-envelope.md../_shared/error-contract.md../_shared/progress-contract.md../_shared/ui-binding.md../_shared/sync-policy.md../_shared/logging-fields.md
Then read:
references/connection-profiles.mdreferences/request-urls.mdreferences/model-catalog.mdreferences/pricing-matrix.mdreferences/capability-matrix.md- the transport for the requested kind
references/logging-contract.mdwhen logging is implementedreferences/model-sync.mdonly for an explicit update or verification task
Read references/hosted-tools.md when provider-hosted tools are requested.
Runtime Rules
- Send the exact documented model ID. This skill currently maintains rules for
gpt-5.6-sol,gpt-5.6-terra,gpt-5.6-luna, andgpt-image-2. - Do not replace a requested model with another model or moving alias.
- Resolve connection profile, request kind, API surface, API version, endpoint kind, full route key, base URL, and request URL before capability lookup.
- Match capabilities by the exact full
RouteKeyrow, including API version and endpoint kind. - Prefer no surface implicitly: use the surface required by the project. Responses and Chat Completions have different payloads and continuation rules.
- Use
gpt-image-2through the direct Image API rows. Responses hosted image generation is a separate tool flow whose main model is a GPT model. - Reject requested fields marked
unsupportedorunknown; do not silently drop or translate them to another surface. - Keep caller-defined functions in
Toolsand provider-hosted tools inHostedTools. - Preserve response item IDs, call IDs, required linkage, encrypted reasoning items, and assistant
phasewhen continuing a documented stateful flow. - Never expose raw chain-of-thought. Map visible summaries to
ReasoningSummaryand opaque state toReasoningItemsor provider metadata. - Store API-key references only. Never write secrets, authorization headers, MCP credentials, or signed URLs into these files or normal logs.
- Use
pricing-matrix.mdfor estimates, including context bands and cache charges.
Request Flow
- Read the requested provider, region, model, and request kind from the host project or user.
- Resolve the connection profile and exact request URL row.
- Resolve the exact catalog and capability rows.
- Validate every optional field before constructing the request.
- Apply the matching transport.
- Normalize the provider response, errors, progress, and logs through the shared contracts.
OpenAI Field Mapping
- Responses
reasoning.effortand Chat Completionsreasoning_effortuse the values recorded in the capability matrix. TextVerbositymaps to Responsestext.verbosityor Chat Completionsverbosityonly where verified.- Responses
reasoning.mode,reasoning.context, andprevious_response_idare Responses-only fields. - Responses structured output maps through
text.format; Chat Completions maps throughresponse_format. CacheKeymaps toprompt_cache_keywhere verified; explicit cache controls require the documented Responses mapping.
Out of Scope
- ChatGPT consumer configuration
- Apps SDK or Agents SDK orchestration
- Realtime/audio integration
- fine-tuning
- model ranking or model permission policy
- non-OpenAI providers