[bbingz]

ELI5

Pressing Option + Delete (⌥⌦ / Alt + Delete) in an Orca terminal now deletes the word to the right, matching standard macOS terminal behavior and preventing stray characters or accidental line clearing in tools like Claude Code.

What Changed

• Before: In terminal panes or Claude Code sessions, pressing Option + Forward-Delete (⌥⌦) sent raw escape sequence \x1b[3;3~ (CSI 3;3~). Shells without this binding would swallow the prefix and print a stray ~, while Claude Code would misinterpret the sequence and delete all text to the end of the line instead of a single word.
• After: Option + Forward-Delete (⌥⌦ on macOS, Alt + Delete on Linux/Windows) translates to \x1bd (ESC d), which readline and CLI tools recognize as forward-kill-word. When the Kitty Keyboard Protocol is active, it defers to xterm so advanced TUIs receive native events.
• Mechanism: In resolveTerminalShortcutAction (src/renderer/src/components/terminal-pane/terminal-shortcut-policy.ts), unified Alt+Backspace and non-numpad Alt+Delete to emit \x1bd for Delete and \x1b\x7f for Backspace.

Why

On macOS and in readline-based shells, Option + Backspace deletes the word to the left (backward-kill-word), and Option + Forward-Delete symmetrically deletes the word to the right (forward-kill-word). Unifying this alongside Alt+Backspace provides expected muscle-memory behavior across terminals (iTerm2, Terminal.app, Ghostty) while keeping cyclomatic complexity low and remaining strictly within line-count limits.

Alternatives considered:

• Letting shells handle raw CSI 3;3~: Standard readline and Zsh ZLE configurations do not bind CSI 3;3~ by default, leading to terminal noise or application misbehavior.
• Handling macOS only: Alt+Delete on Linux and Windows readline shells also expects forward-kill-word (\x1bd), so treating both platforms consistently aligns with the existing Alt+Backspace and Alt+Arrow policies.

Linked Issue

Fixes #21491

Visual Proof

N/A - Keyboard input sequence mapping inside terminal pane; verified with automated unit test suite.

Testing

• Automated tests added:
  • Added dedicated test suite src/renderer/src/components/terminal-pane/terminal-shortcut-forward-delete.test.ts covering macOS, Linux/Windows, Kitty protocol negotiation, numpad exclusion, and modifier isolation.
• Local verification:
  • node config/scripts/check-changed-code-quality.mjs (passed with 0 findings)
  • node config/scripts/check-max-lines-ratchet.mjs (passed with 0 new bypasses; file stays at 293 lines)
  • tsc --noEmit across CLI and Web configurations (all passed)
  • vitest run on terminal-shortcut-forward-delete.test.ts and terminal-shortcut-policy.test.ts (40 passed)

AI Disclosure

Developed with Google Antigravity.

Agent skill upstream boundary

• Not applicable

Checklist

• This PR is small and focused
• I explained what changed and why (ELI5, the user-facing before/after, the mechanism, and why over the alternatives)
• Before/after screenshots or videos attached for UI changes, or N/A with reason
• Self-reviewed for correctness, security, and performance
• Cross-platform, SSH/remote, and path/shortcut impact considered (or N/A)
• pnpm lint, pnpm typecheck, pnpm test, and pnpm build pass (or CI will cover; local preferred)

[coderabbitai]

Understand this PR’s impact

Explore downstream dependencies and potential security impact with Blast Radius.

View blast radius →

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Repository UI

Review profile: CHILL

Plan: Advanced

Run ID: b1aca00f-b951-4ed3-bcbe-7abea67fbcc1

📥 Commits

Reviewing files that changed from the base of the PR and between 7a6d100 and 9b5d240.

📒 Files selected for processing (2)

• src/renderer/src/components/terminal-pane/terminal-shortcut-forward-delete.test.ts
• src/renderer/src/components/terminal-pane/terminal-shortcut-policy.ts

Included review availability: Your plan provides up to 10 included reviews per hour; 3 remain after this review.

---

📝 Walkthrough

Walkthrough

The terminal shortcut policy now maps non-keypad Alt+Delete to \x1bd, while Alt+Backspace remains mapped to \x1b\x7f. The resolver returns null when the Kitty keyboard protocol is active. New tests cover macOS, non-macOS, keypad Delete, plain Delete, Meta+Delete, and Ctrl+Alt+Delete behavior.

Priority: ➖ Normal

Severity of issue fixed: Medium

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 50.00% which is insufficient. The required threshold is 80.00%. Docstring coverage is scoped to functions touched by this diff. Analyzed 2 functions across 2 files.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly and concisely describes the main change: translating Option/Alt + Forward-Delete into the readline forward-kill-word sequence.
Description check	✅ Passed	The description covers the required change, rationale, linked issue, visual-proof exception, testing, AI disclosure, scope boundary, and checklist. The Review and Notes sections are not included, but …
Linked Issues check	✅ Passed	The implementation addresses #21491. For Alt/Option + Delete, excluding numpad Delete, resolveTerminalShortcutAction sends \\x1bd, which maps to readline forward-kill-word and replaces the proble…
Out of Scope Changes check	✅ Passed	The changes stay within #21491. The Backspace branch keeps its existing \\x1b\\x7f behavior, while the shared condition and comment support the new Delete mapping. The added tests verify the requested…

• Fix all pre-merge checks with AI

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[pullfrog]

✅ No new issues found.

Reviewed changes

• Alt/Option + Forward-Delete → forward-kill-word: resolveTerminalShortcutAction now matches non-numpad Delete alongside Backspace in the Alt-modified branch, emitting \x1bd (ESC d) while keeping \x1b\x7f for Alt+Backspace. Kitty-negotiated panes still return null before emitting bytes, so xterm sends native escapes.
• Numpad exclusion: event.code?.startsWith('Numpad') !== true keeps keypad Delete/NumpadDecimal out of the translation, preserving the kitty functional-numpad identity path in resolveTerminalOptionShortcutAction.
• Tests: the new terminal-shortcut-forward-delete.test.ts pins macOS, non-mac, kitty deferral, numpad exclusion, and modifier isolation with exact sendInput byte assertions — they would fail against the pre-PR behavior.

I specifically checked the non-mac half against the local Windows ConPTY hazard the repo already guards for Ctrl+Arrow (Escape→PSReadLine RevertLine). Upstream microsoft/terminal and PSReadLine sources show ConPTY fuses ESC d into a single Alt+d key event, and PSReadLine's default Windows bindings map Alt+d to KillWord — so the unguarded translation here is correct and guarding it would instead send Alt+Delete, which PSReadLine does not bind. This also matches the existing unguarded Alt+Arrow / Alt+Backspace branches. The dashboard preview terminal routes through resolvePreviewShortcutAction, which delegates to this same policy, so there is no parity gap.

  ^{｜ View workflow run ｜ Using DeepSeek Flash (free via Pullfrog for OSS) ｜ 𝕏}

[nwparker]

Thanks @bbingz! Your commit landed via #33296 (kept, plus Co-authored-by): ESC d is sent on macOS only, and full-screen non-agent apps like vim keep the native sequence. Closing as superseded.