Claude Code v2.1.158
Released May 30, 2026 (3 months ago).
220 issues reference this build
20 still open
200 resolved (91%)
"References this build" means the issue body mentions running 2.1.158 (from the
bug-report template). It reflects report volume, not a proven defect count — an issue may mention a
version without being caused by it. First-party data from the anthropics/claude-code tracker.
Top issues reported on v2.1.158
Ranked by community engagement (reactions and comments).
- Recurring error: "The model's tool call could not be parsed (retry also failed)" interrupts sessions
- [BUG] Bash tool calls emitted as raw <invoke> text instead of executing
- [Bug] Parallel tool calls hang on Bash execution errors
- [BUG] Auto mode blocks all Bash actions with endless retries when Opus 4.8 classifier is temporarily unavailable
- [MODEL] Opus 4.8 starts hallucinating results before parallel tasks finish
- [BUG] Automatic parallel-tool-call cancellation is indistinguishable from a user interrupt
- [BUG] Tool-call results empty in live UI then flush late/out-of-order (Bash, Read, MCP, advisor) — macOS, Opus 4.8, parallel batches
- [Bug] Stray token "count" injected before tool calls causes parse failure in Claude Code
- [Bug] Tool-use block markup intermittently corrupted with stray token prefix
- Prompt cache fully re-created after turns with many parallel tool calls (cache_read collapses to system+tools floor) — ~74% of cache writes wasted on Opus 4.8 / v2.1.15x
- [BUG] Enterprise-managed marketplaces in project enabledPlugins are orphaned on first run; only install after a restart
- Claude Code VS Code extension forces 1M context on Pro plan - Windows
- [BUG] Opus 4.8 medium effort spends 46k output tokens on hidden thinking for a simple coding turn
- [BUG] Notion MCP connector (both project .mcp.json and claude.ai OAuth connector) triggers 400 input_schema oneOf/allOf/anyOf error in interactive main loop
- Regression (since 2026-05-29): intermittent "tool call was malformed and could not be parsed" — tool_use block absent on a stop_reason=tool_use turn
- Context hit 1M token limit with no auto-compaction and no recovery path
- [Bug] Usage credits required for 1M context on Max 5x plan — standard model selected, Cowork + Claude Code
- [BUG] Repeated Bash tool-call loop during git workflow consumed session limit
- [BUG] typescript-lsp pushes stale diagnostics on 2.1.158 (post-#17979) in TS project-references/composite workspaces
- [Bug] Claude 4.8 model stuck in infinite loop executing spurious bash commands
- [MODEL] Claude confabulated a prompt-injection payload in a file before the file-read result returned
- [Bug] Claude Opus 4.8 emits malformed tool calls with parsing errors
- Harness silently executes duplicated parallel tool_use blocks: subagent fan-out runs N× the intended count (6 → 24)
- [BUG] Desktop file viewer blocks files in permissions.additionalDirectories ("This file is outside the working directory") — regression
- [BUG] Windows + Git Bash: tengu_cobalt_ridge experiment silently force-enables the PowerShell tool & "Shell: PowerShell", causing wrong-shell retries (token waste)
- Server-side tool (server_tool_use) emitted but no result returned; response reports stop_reason: end_turn
- [BUG] Agent View (Linux): mouse wheel sends arrow keys instead of scrolling attached session — works on macOS, fails on all Linux terminals
- [Bug] Opus 4.8 responses not displayed after tool use — quota consumed with no output
- Stray "court" token emitted immediately before tool-use <invoke> tag corrupts the tool call (Opus 4.8, Claude Code 2.1.158)
- [Bug] Parallel tool calls cancel all siblings on single error, which appears to cause model confusion
Changelog
What's changed
- Auto mode is now available on Bedrock, Vertex, and Foundry for Opus 4.7 and Opus 4.8. Opt in by setting
CLAUDE_CODE_ENABLE_AUTO_MODE=1