Pattern Proposal: Provider-Agnostic LLM Adapter
Summary
Encapsulates disparate model endpoints behind a uniform interface supporting production and mock execution modes.
Candidate Pattern
- ID:
provider-agnostic-llm-adapter - Classification: intelligence
- Maturity: established
- Type: implemented
Observed Signals
No specific signals listed.
Evidence
Supporting Evidence
- src/cas_runner/llm/litellm_client.py (LiteLLMClient) - Production LLM integration wrapping litellm library
- src/cas_runner/llm/mock_client.py (MockLLMClient) - Deterministic test double for agent logic verification
Scope Assessment
- Evidence Strength: high
- Confidence: high
- Applies to: ideas-to-life-cas
Rationale
- Intent: Decouple architectural inference steps from concrete model providers to facilitate testing and multi-model flexibility.
- Context: Used across cas_runner whenever agent reasoning, delta evaluation, or text merging is required.
- Force: Need to test agent logic deterministically without API costs or network flakiness
- Force: Need to route requests across diverse cloud LLM vendors
Traceability
Artifact source: proposed-patterns