Repository navigation
Conversation
Prints the git toplevel's name, or repo@worktree for a linked worktree (format overridable via tabname.worktree-format). Nothing and exit 1 outside a work tree. A script so the claude() wrapper and an Alfred switcher can share one answer. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…Code's own title iTerm2's Session Status row is named by the session name, which an OSC 0 title sets. Claude Code writes its rename/AI/haiku title there, so rows read as conversation topics. The wrapper now exports CLAUDE_CODE_DISABLE_TERMINAL_TITLE=1 (off-spellings normalised to unset, so one launch can get Claude's titles back) and writes one title itself before exec: the tabname label, prefixed by WORKTABS_GROUP, or the WORKSET name. Only an attended launch gets the bytes; the attended test is hoisted out of the second-session question and shared. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The claude() wrapper now titles the tab with the repo, and the row name is that title, so prefixing each Session Status detail with the cwd's basename only spent room. The detail carries the status alone. A session launched without the wrapper keeps Claude's own title as its row name. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…ude's own off iTerm2 3.7.5+'s Claude Code workgroup renames the session to "Chat" on job start, so a title written before exec is overwritten within a second. The cc-status shim will name the session after that; the wrapper keeps the CLAUDE_CODE_DISABLE_TERMINAL_TITLE default so Claude Code's own OSC 0 titles do not compete, and loses the OSC 0 write and its pty tests. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
On iTerm2 3.7.5+ the Claude Code workgroup renames the session to "Chat" when the job starts, so every Session Status row read "Chat (claude)". The shim now runs it2 session set-name with [WORKTABS_GROUP] and WORKSET or tabname of the cwd, on a session's first UserPromptSubmit (marked by <session_id>.named) and on every Stop and StopFailure, never on SessionStart. It is an in-memory override; no iTerm2 pref is written. The rename runs after cc-status and before the paint, so the paint stays the last it2 call of a turn end. explain prints name= and rename=. The sh gate now sends every Stop and StopFailure to python3. tests/cc-status-shim.bats gains a tabname stub and 11 rename tests. Counts changed for Stops carrying a cwd: the seat Stop over an active seat and the question Stop with a quiet task go from 1 it2 call to 2; the seat-not-among-tasks, stale-state-file, statement-ending, active-task question and no-question-mark Stops go from "no it2 call" to exactly the rename (rename_only). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
… title Record why rows read "Chat (claude)" on iTerm2 3.7.5+ (the Claude Code workgroup renames the session on job start), that the cc-status shim renames it to the repo through it2 session set-name, and that no pref in the tracked iTerm2 plist is part of the mechanism. Also tabname's contract, the wrapper's CLAUDE_CODE_DISABLE_TERMINAL_TITLE default and its costs, and the shim's new per-turn python3 cost. The shared memory counts five wrapper defaults. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…ET needs no tabname The .named marker is now written on every first UserPromptSubmit, not only after a rename, so a session outside a repo or without tabname costs one python start rather than one per prompt. session_title() checks $WORKSET before looking for tabname: a working set names the row even with no tabname installed and outside a repo. explain gains marker=present|none. Tests: the "outside a repo, without tabname, or without it2" case stays; a new test asserts the marker for not-a-repo and no-tabname and that the second prompt skips python (no it2 call, no tabname call, payload fed unchanged); without it2 the sh gate passes before python so there is no marker. Two WORKSET tests and one explain marker test added; the explain test asserts marker=. No existing count assertion changed. CLAUDE.md: the wording that said no tabname always means no rename now excepts WORKSET, and notes the marker. The "uncommitted" plist sentence is step 8's. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…his clone holds The tab-name paragraph described "the plist's uncommitted Workgroups blob", which is a fact about one Mac's working tree. Say instead that the Claude Code onboarding writes the workgroup and its Enter/Exit triggers into the plist itself. The step-7 wording fix ($WORKSET needs no tabname) was already in. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…nal-title is false CLAUDE_CODE_DISABLE_TERMINAL_TITLE=1 was exported on every machine, but it only pays where the cc-status shim, iTerm2 3.7.5+ and the integration are all present. Gate it on machine-local git config, default off, so a public-only clone or a Linux box keeps Claude's own titles. claude-doctor gains a titles: line. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
iTerm2 Session Status rows, and the tab names beside them, now show the repo instead of Claude Code's conversation titles.
Why the rows said "Chat (claude)"
On iTerm2 3.7.5 and later, the Claude Code workgroup renames a session to the workgroup root's display name ("Chat") when
claudestarts. iTerm2 then adds the job, so the result is "Chat (claude)". OSC 0 titles no longer name that session.What changes
bin/tabnameprints the label for a directory: the git toplevel's name, or the repo then an@then the worktree name for a linked worktree, with the format set bytabname.worktree-format. Outside a work tree it exits 1 and prints nothing. The Alfred session switcher planned next will use it too.it2 session set-nameon the first UserPromptSubmit and on every Stop, so a lost rename heals at the next turn end. The name is[<WORKTABS_GROUP>] <WORKSET or tabname>, it lives in memory only, and no iTerm2 preference is involved. The detail line no longer repeats the repo.claude()exportsCLAUDE_CODE_DISABLE_TERMINAL_TITLE=1only where machine-local git config saysclaude.terminal-title = false, and the overlay sets that on the Mac whose hooks run the shim. With no config, Claude's own titles are untouched, so a public-only clone or a Linux box loses nothing.claude-doctorgains atitles:line.Landing order
Merge the personal overlay PR (the
claude.terminal-titleline) first, then this one, then rundotfiles.Tests
The full suite passes: 713 of 713 ok. On this Mac the shim tests need
/usr/binahead of the asdfpython3shim under a sandboxed HOME, which is not caused by this branch.Not verified yet
claudetab should read<repo> (claude)after the first prompt./resumetitles after a week with the opt-out on.Chat (-zsh)tabs need a one-timeit2 session set-name "$(tabname)".🤖 Generated with Claude Code