Release v2.1.86 · marckrenn/claude-code-changelog

2.1.86

Highlights:

Added X-Claude-Code-Session-Id header to requests so proxies can group sessions without parsing bodies
Added debug-before-escalate rule: require on-call diagnosis/fix attempts before escalation to cut handoffs
Removed requirement for Bash command descriptions, allowing concise command-only responses for easier parsing

Added X-Claude-Code-Session-Id header to API requests so proxies can aggregate requests by session without parsing the body
Added .jj and .sl to VCS directory exclusion lists so Grep and file autocomplete don't descend into Jujutsu or Sapling metadata
Fixed --resume failing with "tool_use ids were found without tool_result blocks" on sessions created before v2.1.85
Fixed Write/Edit/Read failing on files outside the project root (e.g., ~/.claude/CLAUDE.md) when conditional skills or rules are configured
Fixed unnecessary config disk writes on every skill invocation that could cause performance issues and config corruption on Windows
Fixed potential out-of-memory crash when using /feedback on very long sessions with large transcript files
Fixed --bare mode dropping MCP tools in interactive sessions and silently discarding messages enqueued mid-turn
Fixed the c shortcut copying only ~20 characters of the OAuth login URL instead of the full URL
Fixed masked input (e.g., OAuth code paste) leaking the start of the token when wrapping across multiple lines on narrow terminals
Fixed official marketplace plugin scripts failing with "Permission denied" on macOS/Linux since v2.1.83
Fixed statusline showing another session's model when running multiple Claude Code instances and using /model in one of them
Fixed scroll not following new messages after wheel scroll or click-to-select at the bottom of a long conversation
… +14 more CLI changelog entries

Time since 2.1.85 release: 23h 34m 30s
Bundle file size: +47.8 kB (+0.4%)
Est. LOC (prettified): +1,420 (+0.4%)
Prompt files: +4 (+0.1%)
Prompt tokens: +821 (+0.2%)

Prompt token mix by kind:

Files:

Notable changes:

Claude’s Skill guidance got shortened: update-config/loop/schedule descriptions lost details/examples, and claude-api “DO NOT TRIGGER” exclusions were cut. This can change when Claude decides to invoke these skills and reduces specificity for correct routing.
Claude is now instructed to investigate failures first: read errors, check assumptions, try a targeted fix; avoid blind repeat retries but also avoid abandoning a viable approach after one failure. AskUserQuestion is framed as escalation only after real investigation.
Claude’s “avoid premature abstraction” rule now adds a second constraint: avoid half-finished implementations. Net effect: Claude should still keep designs minimal, but is nudged to fully implement what the task requires instead of stopping at partial solutions.
Claude no longer has an explicit Bash guideline to write a clear, user-facing description of each command (with extra context for complex pipelines/flags). This can reduce how much Claude explains Bash actions before running them.

Links: