# Known flaky E2E tests — LLM-dependent, timeout on slow CI runners or when API key is unavailable.
# These tests pass locally with GEMINI_API_KEY set.
# Format: file:line — description
#
# To run ALL tests including flaky: npx playwright test
# To skip flaky in CI: set SKIP_KNOWN_FLAKY=true

# LLM response timeouts — require live Gemini API
ai-tools.spec.ts:104 — greeting receives a response (LLM timeout)
unified-console.spec.ts:56 — user can send a message and receive a response (LLM timeout)
rag-chat.spec.ts:27 — user can send a prompt and receive streaming response (LLM timeout)
rag-chat.spec.ts:48 — RAG-specific prompt receives a response (LLM timeout)
rag-chat.spec.ts:58 — transport-specific prompt receives a response (LLM timeout)
rag-chat.spec.ts:68 — multi-turn conversation preserves history (LLM timeout)

# Java integration tests — CI runner timing
ChatIntegrationTest.longPollingTransportConnectsAndReceivesMessage — long-polling timeout on slow CI
