Runs the Olla sticky session integration test end-to-end across all provider-scoped routes. Trigger when the user asks to: verify sticky sessions work, run the sticky session integration test, test provider-route affinity, or check whether the providerProxyHandler bug fix is holding. Delegable to Sonnet - does not require Opus.
End-to-end proof that Claude Code works turn-by-turn through Olla's Anthropic surface against a real Ollama backend. Boots a fresh Olla in front of a user-supplied Ollama endpoint, then exercises both legs - native Anthropic passthrough and forced OpenAI<->Anthropic translation - with two tiers: Tier-A direct-wire protocol assertions (curl + jq against Olla) and Tier-B a real headless Claude Code run implementing a fixed feature in a fresh clone of thushan/smash. Use to validate the Anthropic translation/passthrough layer under genuine Claude Code traffic. Needs a working Ollama and the claude CLI; not CI-safe (real LLM, real network). Runs on Sonnet.
Full automated validation of Olla via parallel agents against mock backends (no Docker, no real LLMs - CI/CD safe). Gates on make ready first, then boots an ollamock fleet + Olla and fans out per-area validation agents covering core routing, OpenAI/Anthropic surfaces, observability, health/recovery and failure handling. Use as the pre-release gate or after major changes. Trigger when the user asks to validate Olla, run the validation suite, run regression/nightly validation, or gate a release. Runs on Sonnet regardless of the session model; area agents are balanced between Sonnet and Haiku for token efficiency.