Repository Issue Activity (beta)

#5906 Clawpatch automated code audit: 666 findings across 295 features

#5911 passthrough inference provider creates httpx client per request

#6343 remote::passthrough inference provider broken on openai >= 2.34.0 — all requests return 500

#6245 Proposal for gradient text based logo for OGX website navbar

#3769 startup crash when config registers a model and provider returns an exception

#4697 Responses API reasoning summaries from providers that provide them

#2912 configure vector-io provider with an embedding model

#5846 OpenAI -pro models fail via Responses API: OGX should forward to OpenAI /v1/responses instead of /v1/chat/completions

#2296 Road to v1

#5608 docs: remove todo as Neural Reranking has been already implemented

#6263 Tool outputs (websearch, file_search) fed unsanitized into LLM context — indirect prompt injection

#6242 Responses API `file_search` does not populate `file_citation` annotations

#5478 proposal: Package-per-provider using uv workspaces and Python entry points

#5588 bug: reserved __authenticated_user key in provider data can corrupt auth context and trigger 500s

#6018 User message dedup drops messages when using previous_response_id

#6020 Structured output json_schema.strict field stripped, causing Anthropic 400 error

#5601 docs: versions.html is outdated and links are broken

#5755 OpenAI Vector Store File Ingestion Does Not Pass `input_type` for NVIDIA Asymmetric Embedding Models

#5892 meta: Containers API + ContainerRuntime Provider

#5910 Expose httpx connection pool config for inference providers

#5230 feat: support Bedrock models beyond the Mantle API (e.g. Amazon Nova, native Converse API)

#5904 Migrate Interactions API to Google's new steps schema (deadline: May 26, 2026)

#4516 Database Schema Management CLI Proposal

#5008 Vector stores return "not found" immediately after creation when using multiple server workers

#5880 testing: unconditional `import ollama` in providers/__init__.py forces ollama dep for all test suites