quickstart-resources

Motivation and Context PR #141 fixed the Anthropic Messages API tool-use handling in the Python client (). The TypeScript client () carried the same bugs, but they stayed latent because the smoke test's mock MCP server returns an empty tool list, so the tool-use branch never executes. The original TS is not spec-compliant: it sends each tool result back as a plain-text message (no block, no ), never appends the assistant turn carrying the block, drops from follow-up calls, and has no loop for chained tool calls. The observable symptom is not a hard API error — the Messages API tolerates the malformed shape (it accepts the result as plain user text). Instead the bug shows up as silently incorrect answers and broken tool chaining. Verified against the real Messages API on the pristine build with a query that needs two sequential tool calls ("Compare the weather in NYC and San Francisco"): the model called for NYC, then — because was dropped on the follow-up — could not call it again for SF, so SF was never fetched and the model fabricated the San Francisco half of the comparison. Separately, crashes with on stdin EOF (Ctrl-D) / SIGINT (Ctrl-C). This PR mirrors #141 for TypeScript: Reply to each with a proper block carrying , and append the assistant's full content (preserving the block) instead of sending the result as a plain user string. Restructure into a turn-level loop so parallel tool calls in a single response and chained tool calls across turns both work; pass on every follow-up request so the model can keep using them. Add a cap with a notice to prevent unbounded tool-use loops. Fix to exit cleanly on EOF (Ctrl-D) and SIGINT (Ctrl-C). Node's does not reject on stdin EOF on its own (it just hangs / then throws ), so an wired to the readline event unblocks it. How Has This Been Tested? Tested side-by-side against a pristine build to confirm the fix changes real behavior: [x] passes locally (5/5 green). [x] Structural verification via a fake endpoint that records requests: confirmed the follow-up call sends with , keeps the array, and makes a third call when the model chains another . [x] Manual, real against : Single tool call ("What's the weather in NYC?") — model receives the tool result and produces a final answer. Chained tool calls ("Compare the weather in NYC and San Francisco.") — both calls execute and the final answer is built from real fetched data for both cities. On the pristine build the same query fetched NYC only and fabricated SF. [x] cap path verified by lowering the constant locally and confirming the notice appears. [x] EOF (Ctrl-D) at the prompt exits cleanly (exit 0) instead of crashing with as does. Breaking Changes N/a Types of changes [x] Bug fix (non-breaking change which fixes an issue) Checklist [x] I have read the MCP Documentation [x] My code follows the repository's style guidelines [x] New and existing tests pass locally [x] I have added appropriate error handling Additional context Mirrors #141 (which fixed the same issue, originally #28, for the Python client). 🦉 Implementation and verification assisted by Claude Code. 🤖 Generated with Claude Code

Is your feature request related to a problem? Please describe. A clear and concise description of what the problem is. Ex. I'm always frustrated when [...] Describe the solution you'd like A clear and concise description of what you want to happen. Describe alternatives you've considered A clear and concise description of any alternative solutions or features you've considered. Additional context Add any other context or screenshots about the feature request here. Request Id: 51a69df9-34cd-4645-b4d0-9473a2482700 Correlation Id: 9db70615-8e3e-4f43-949b-fa1b2d833baa Timestamp: 2026-06-12T23:41:21Z Message: AADSTS900561: The endpoint only accepts POST requests. Received a GET request.