Contributing to AI tooling and agent infrastructure this week.
- openai/openai-agents-js#1090 - tightened raw model stream event typings so SDK consumers get precise OpenAI stream event types instead of
any - modelcontextprotocol/typescript-sdk#1674 - fixed missing-tool failures so they surface as proper JSON-RPC errors instead of wrapped tool results
- vercel/ai#13376 - fixed SSRF validation so safe inline
data:image URLs are accepted without weakening network URL protections - continuedev/continue#11352 - documented the full Continue CLI slash command set in TUI mode
- continuedev/continue#11353 - added a new CLI hooks guide and clarified which hook events the CLI emits today
- continuedev/continue#11354 - improved SSL certificate troubleshooting and self-hosting guidance
- continuedev/continue#11356 - fixed broken model configuration docs links in the GUI and added a regression test
- continuedev/continue#11357 - documented the AI SDK toggle and background-job polling flow in the CLI docs
- TypeScript developer tooling
- agent workflows and infrastructure
- AI tooling docs, DX, and small safe OSS fixes
- openai/openai-agents-js#279 - linked a focused typing fix back to the original issue
- modelcontextprotocol/typescript-sdk#1510 - linked a protocol-compliance fix back to the original issue
- vercel/ai#13354 - linked the new inline
data:URL fix back to the original bug report - modelcontextprotocol/typescript-sdk#1639 - confirmed the current npm/package-name mismatch with fresh registry and repo evidence
- modelcontextprotocol/typescript-sdk#509 - clarified that the posted repro uses the wrong MCP request shape and skips initialization