| name | llm-provider-integration |
| description | Use when working on GitHub Copilot, OpenAI-compatible providers, model discovery, connection testing, provider config, env resolution, streaming behavior, or model fallback logic. |
| triggers | ["\\bllm\\b","\\bprovider\\b","\\bmodel.*config\\b","\\bstreaming\\b","\\bopenai\\b","\\bcopilot.*provid\\b"] |
LLM provider integration
Use this skill when the backend work touches provider setup, model resolution, connection handling, or provider-specific runtime quirks.
Typical triggers:
- provider setup or switching is failing
- model discovery is wrong or incomplete
- API keys or env resolution behave incorrectly
- streaming responses stall or break
- provider-specific parameter fallbacks are needed
- connection testing or init setup needs changes
Focus areas
packages/core/src/llm/index.ts
packages/service/src/commands/provider.ts
packages/service/src/commands/models.ts
packages/service/src/commands/test-connection.ts
packages/service/src/commands/init.ts
Workflow
- Identify whether the issue is config resolution, credential loading, model selection, request shaping, or streaming.
- Prefer official provider documentation and current model guidance before changing endpoints, request parameters, or model assumptions.
- Keep provider abstraction in
packages/core and command/setup flow in packages/service.
- Test both happy paths and provider-specific fallback paths.
- Keep logs and diagnostics useful enough to debug real provider failures.
Guardrails
- do not hardcode one provider's assumptions into the abstraction
- do not leak secrets into logs
- keep
.env and config handling explicit and auditable
- prefer targeted fallbacks over provider-specific branching everywhere
Successful outcome
- provider integration remains flexible and reliable
- setup, testing, and runtime usage behave consistently
- model and credential handling stay understandable