Issues

• [BUG] Anthropic API Error: Opus 5.5 Safeguards Incorrectly Flagging Biology Prompt

  bug
  anthropics/claude-code#106995 · Filter by author IricsDo opened on Nov 3, 2026

• [BUG] VS Code: effort is set to high everywhere and logged as high, but sessions reason at a low level

  area:modelbugplatform:linuxplatform:vscode
  anthropics/claude-code#106994 · Filter by author ANSHAY opened on Nov 3, 2026

• Model bundles unrelated destructive and routine actions into one shell command, so one permission denial blocks all of them

  area:basharea:modelarea:permissionsbug
  anthropics/claude-code#106993 · Filter by author Nahuel990 opened on Nov 3, 2026

• OAuth discovery builds malformed URL for remote MCP server (missing / before .well-known)

  area:autharea:mcparea:pluginsbugplatform:windows
  anthropics/claude-code#106992 · Filter by author gruposfocales opened on Nov 3, 2026

• [BUG] Desktop (Windows): sending a message mid-turn creates an unrequested fork; "Start locally" task chip opens a same-folder child session without saying so

  area:desktopbugplatform:windows
  anthropics/claude-code#106991 · Filter by author doubleonoseven opened on Nov 3, 2026

• Model adds more paid cloud steps after the user explicitly said to stop spending

  area:costarea:modelbug
  anthropics/claude-code#106990 · Filter by author Nahuel990 opened on Nov 3, 2026

• Model runs test suites concurrently against shared state and reports the resulting failures as possible regressions

  area:modelbug
  anthropics/claude-code#106989 · Filter by author Nahuel990 opened on Nov 3, 2026

• Model ignores a standing instruction to sweep the whole affected area after an external bug report

  area:modelbugmemory
  anthropics/claude-code#106988 · Filter by author Nahuel990 opened on Nov 3, 2026

• Model calls a fix correct and complete without enumerating all the ways the bug can happen

  area:modelbug
  anthropics/claude-code#106987 · Filter by author Nahuel990 opened on Nov 3, 2026

• Model keeps giving long answers after repeated requests for one-line replies

  area:modelbug
  anthropics/claude-code#106986 · Filter by author Nahuel990 opened on Nov 3, 2026

• Model reports an empty tool result as fact without cross-checking

  area:modelbug
  anthropics/claude-code#106985 · Filter by author Nahuel990 opened on Nov 3, 2026

• Model does not reload persistent memory after context compaction, despite project instructions requiring it

  area:corearea:modelbug
  anthropics/claude-code#106984 · Filter by author Nahuel990 opened on Nov 3, 2026

• Model ends tasks with 'not applied' lists instead of finishing or flagging blockers up front

  area:modelbug
  anthropics/claude-code#106983 · Filter by author Nahuel990 opened on Nov 3, 2026

• Model abandons an unfinished task and switches to another without saying the first is blocked

  area:modelbug
  anthropics/claude-code#106982 · Filter by author Nahuel990 opened on Nov 3, 2026

• Model starts implementing changes without being asked, despite explicit zero-autonomy instructions in memory

  area:modelbugmemory
  anthropics/claude-code#106981 · Filter by author Nahuel990 opened on Nov 3, 2026

• Model writes a regression test that passes without the fix and presents it as coverage

  area:modelbug
  anthropics/claude-code#106980 · Filter by author Nahuel990 opened on Nov 3, 2026

• Model's generated shell scripts break on CLI argument parsing and invalid resource definitions

  area:modelbug
  anthropics/claude-code#106979 · Filter by author Nahuel990 opened on Nov 3, 2026

• Model writes cloud capture scripts without validating their preconditions, costing the user reruns and money

  area:modelbug
  anthropics/claude-code#106978 · Filter by author Nahuel990 opened on Nov 3, 2026

• Model flips its verdict between two equivalent pull requests repeatedly without new evidence

  area:modelbug
  anthropics/claude-code#106976 · Filter by author Nahuel990 opened on Nov 3, 2026

• Model accepts AI-generated pull requests after reading only the diff, missing security gaps in sibling code paths

  area:modelbug
  anthropics/claude-code#106975 · Filter by author Nahuel990 opened on Nov 3, 2026

• [Bug] Session names leak into unrelated sessions' transcripts (ai-title/custom-title)

  area:corebugplatform:windows
  anthropics/claude-code#106974 · Filter by author MoleDJ opened on Nov 3, 2026

• [Feature Request] Reduce autonomous behavior to improve comprehension and collaboration

  area:modelenhancementplatform:macos
  anthropics/claude-code#106973 · Filter by author dylmiller opened on Nov 3, 2026

• [Bug] Opus 5.5 responds in English despite language: Korean setting in long sessions

  area:modelbugplatform:macos
  anthropics/claude-code#106971 · Filter by author loopy-tatum opened on Nov 3, 2026

• [GitHub integration]

  buggithub-integrationneeds-info
  anthropics/claude-code#106970 · Filter by author Mohamed17Amer opened on Nov 3, 2026

• [Bug] Legitimate SDK bridge work for image editors incorrectly flagged as cybersecurity

  area:modelbugplatform:windows
  anthropics/claude-code#106969 · Filter by author CloudyTabzy opened on Nov 3, 2026