Pattern Proposal: LLM provider adapter and test double port

Summary

LiteLLMClient and MockLLMClient provide separate provider clients under an llm package, abstracting LLM transport from business logic.

Candidate Pattern

• ID: llm-provider-adapter-port
• Classification: architectural
• Maturity: emerging
• Type: emerging

Observed Signals

No specific signals listed.

Evidence

Supporting Evidence

• src/cas_runner/llm/litellm_client.py (LiteLLMClient) - Production LLM provider client.
• src/cas_runner/llm/mock_client.py (MockLLMClient) - Test double for provider interaction.
• pyproject.toml - Declares litellm>=1.83.14 as a dependency.

Scope Assessment

• Evidence Strength: high
• Confidence: high
• Applies to: ideas-to-life-cas

Rationale

• Intent: Decouple application and governance logic from a concrete LLM provider and enable deterministic testing.
• Context: The project depends on litellm and likely needs both real provider calls and fake responses for tests or dry runs.
• Force: Real provider calls are expensive and nondeterministic
• Force: Tests need stable provider behavior
• Force: Provider APIs may change faster than domain logic

Traceability

Artifact source: proposed-patterns