Error listings

All error pages - page 21

AI Coding Tools Claude Code Updated May 21, 2026

Claude Code MCP STDIO subprocess reaped mid-conversation without shutdown signal

Users running Claude Code with MCP STDIO-based tools experience server processes getting OS-reaped during conversations, then respawned without graceful shutdown, causing hung connections and tool failures. Includes evidence for Claude Code troubleshooting demand.

MCP STDIO subprocess reaped + respawned mid-conversation, no shutdown signal to the server
AI Coding Tools Claude Code Updated May 21, 2026

Claude Code Opus 4.7 tool call parsing fails with unrecoverable retry loop

Claude Code Pro/Max users running Opus 4.7 encounter tool call parsing failures where retries also fail, leaving agents stuck. Regression started May 20, 2026. Labeled 'has repro' with detailed reproduction steps. Includes evidence for Claude Code troubleshooting demand.

could not be parsed (retry also failed) — Opus 4.7 tool calls fail repeatedly starting 2026-05-20
AI Coding Tools Claude Code Updated May 21, 2026

Claude Code remote routines lose MCP connector permissions every run

Users running scheduled webhooks/retrieval-routines in Claude Code web UI lose all MCP connector permissions on each invocation due to ephemeral container recreation, requiring manual permission grant every time. Includes evidence for Claude Code troubleshooting demand.

Remote routines lose MCP connector permissions every run — tools blocked in ephemeral container
AI Coding Tools Claude Code Updated May 21, 2026

Claude Code Shopify MCP token-expired loop after multi-shop switch

Merchants using Claude Code with Shopify MCP connector switch between multiple shop instances and get stuck in perpetual token-expired error loops that cannot be resolved without restarting the entire session. Includes evidence for Claude Code troubleshooting demand.

token-expired loop after multi-shop switch, unrecoverable from client side
OpenAI API OpenAI API Updated May 21, 2026

OpenAI Responses API background mode fails to surface typed exceptions for failed runs

Developers using OpenAI Python SDK background responses API encounter opaque failure payloads instead of typed RetryableError/InternalServerError exceptions, breaking production retry/backoff logic. Includes evidence for OpenAI API troubleshooting demand.

Status='failed', error.code not mapped to SDK exception class; HTTP poll returns 200 OK but no typed exception raised
OpenAI API Paperclip (AutoGen-like Agent Framework) Updated May 21, 2026

Agent Runtime Dies on 429 Rate-Limit with No Recovery Path

Production agents running Anthropic/OpenAI get permanently killed by transient 429 errors; users need automatic retry/fallback chain instead of manual intervention Includes evidence for Paperclip (AutoGen-like Agent Framework) troubleshooting demand.

stopReason=error; agent transitions to error state after provider returns 429/rate-limit; no retry or model fallback occurs
AI Coding Tools Claude Code Updated May 21, 2026

Claude Code MCP Permission Denied Keeps Prompts Despite Repeated Authorizations

Claude Code users configuring MCP servers face persistent permission denied errors despite granting allow permissions; need fix for Claude Code MCP authorization behavior Includes evidence for Claude Code troubleshooting demand.

Claude Code keeps asking for permission even with 'Allow' setting; MCP plugin reports 'Permission denied' repeatedly
LiteLLM LiteLLM Updated May 21, 2026

LiteLLM Silently Swallows Mid-Stream 429 Rate-Limit Errors During Streaming

Users deploying LiteLLM proxy with Vertex AI/Gemini streaming see silent failures where 429 errors are hidden — stream appears successful but response is truncated. Need to detect and handle mid-stream rate limits properly. Includes evidence for LiteLLM troubleshooting demand.

LiteLLM terminates stream with finish_reason="stop" instead of raising 429 RESOURCE_EXHAUSTED; client receives [DONE] with no error indication
LiteLLM LiteLLM Updated May 21, 2026

LiteLLM Timeout Silently Ignored for Bedrock and Vertex AI Providers

Users of LiteLLM proxy with AWS Bedrock or Vertex AI hit timeouts but the timeout config is silently ignored, leaving requests hanging for 600s before surfacing to user Includes evidence for LiteLLM troubleshooting demand.

litellm.Timeout: Connection timed out. Timeout passed=600.0, time taken=600.006 seconds; error returned directly to user
OpenAI API OpenAI API / Codex Updated May 21, 2026

OpenAI Codex App 429 Rate Limit Exceeded — Retry Loop Blocks Paid Plus Users

Paid Codex Plus subscriber gets stuck in infinite retry loop when hitting DDoS-rate limits; needs solution to stop repeated 429 blocking usage Includes evidence for OpenAI API / Codex troubleshooting demand.

exceeded retry limit, last status: 429 Too Many Requests
Anthropic API Anthropic API Updated May 21, 2026

Anthropic Vertex AI 413 Payload Too Large: Incorrect rejection of requests well under documented 32MB limit

Developer sends ~2MB document payload to Claude Sonnet 4 via Vertex AI and gets 413 error despite being well under the documented 32MB limit — debugging difficult because error misreports the cause Includes evidence for Anthropic API troubleshooting demand.

Error code: 413 - Request exceeds the maximum allowed number of bytes. The maximum request size is 32 MB (but payloads of ~2MB are rejected)
LiteLLM LiteLLM Updated May 21, 2026

LiteLLM proxy buffers Anthropic streamed responses instead of sending deltas incrementally

Production LiteLLM proxy user expects incremental streaming for Anthropic model calls but receives all content in a single buffered batch — breaks real-time chat UX Includes evidence for LiteLLM troubleshooting demand.

Proxied Anthropic Response is Buffered: response is sent in one large batch despite stream=true setting; works correctly for OpenAI models under same config
Deployment Vercel Updated May 21, 2026

Vercel CLI 54.2.0 breaks SvelteKit route resolution for /admin/* paths

Developer upgrades Vercel CLI and discovers SvelteKit admin routes stop working — routing table changed between CLI versions causing deployment failures Includes evidence for Vercel troubleshooting demand.

CLI 54.2.0 breaks SvelteKit route resolution for /admin/* routes
Deployment Vercel Updated May 21, 2026

Vercel static-build production Vite build produces import.meta.env.DEV=true after vite.resolveConfig pre-scan

Developer deploying Vite app to Vercel finds production build incorrectly has import.meta.env.DEV=true instead of false — causes dev-only code to ship to production Includes evidence for Vercel troubleshooting demand.

Production Vite build produces import.meta.env.DEV=true after vite.resolveConfig pre-scan (@vercel/static-build@2.9.28, CLI 54.2.0)
AI Coding Tools Claude Code Updated May 21, 2026

Claude Code Pro to Max Plan Upgrade Fails Silently — PaymentIntent Voided Before Bank Contact

User cannot upgrade from Claude Pro ($20/mo) to Claude Max plan; payment fails without any bank notification or OTP prompt;怀疑 Anthropic 服务端 billing state machine bug Includes evidence for Claude Code troubleshooting demand.

[BUG] Pro → Max upgrade fails silently — no bank notification, PaymentIntent voided before bank is contacted
AI Coding Tools Claude Code Updated May 21, 2026

Claude Code Session Burns Tokens Without Interaction — File Does Not Exist Error Repeating

Claude Code automatically consumes 15% of user's 5-hour quota on session startup without any user input; repeated FileDoesNotExist errors suggest background process or config scanning bug wasting paid API tokens Includes evidence for Claude Code troubleshooting demand.

Error: File does not exist. Note: your current working directory is /Users/edgar/Desktop/admin-app-ios at call (/$bunfs/root/src/entrypoints/cli.js:4299:9752)