fix(cli): Always process @file references in interactive mode

[edlsh]

Summary

Fixes #14919

When users create prompts using the external editor (Ctrl+X), @file references were not being automatically loaded. Instead, the LLM had to initiate file reads itself, wasting tokens.

Root Cause

Interactive mode used isAtCommand() as a guard before calling handleAtCommand(). This heuristic (query.startsWith('@') || /\s@/.test(query)) failed in certain cases—especially for multi-line prompts from the external editor where @file might not be at the start or preceded by whitespace.

Non-interactive mode already calls handleAtCommand() unconditionally, so the fix is to match that behavior.

Changes

1. useGeminiStream.ts: Remove the isAtCommand() conditional guard—now always calls handleAtCommand() to match non-interactive behavior
2. atCommandProcessor.ts: Add a fast-path early exit (if (!query.includes('@'))) to avoid unnecessary parsing when no @ symbol exists
3. atCommandProcessor.test.ts: Add 3 new tests for edge cases:
   • Prompts without @ symbol (verifies fast path)
   • Email addresses (verifies they don't trigger file loading)
   • Multi-line prompts with @file on non-first line

Testing

• All 47 tests in atCommandProcessor.test.ts pass
• Full npm run preflight passes (build, typecheck, tests)
• Manual verification recommended with external editor workflow

[gemini-code-assist[bot]]

Summary of Changes

Hello @edlsh, I'm Gemini Code Assist[^1]! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

This pull request resolves a bug that prevented @file references from being correctly processed when users composed prompts using the external editor in interactive mode. By ensuring that @-commands are always evaluated, regardless of the input method, the change standardizes behavior across different modes and optimizes token usage. Additionally, a performance improvement has been introduced for queries that do not contain @ symbols, alongside new tests to validate these scenarios.

Highlights

• Consistent @file processing: The pull request fixes an issue where @file references were not automatically loaded when prompts were created using the external editor in interactive mode, leading to wasted LLM tokens. The fix ensures handleAtCommand() is always called, aligning interactive mode behavior with non-interactive mode.
• Performance Optimization: A fast-path early exit (if (!query.includes('@'))) has been added to atCommandProcessor.ts. This prevents unnecessary parsing when a query does not contain an @ symbol, improving efficiency.
• Enhanced Test Coverage: Three new tests have been added to atCommandProcessor.test.ts to cover edge cases, including prompts without @ symbols, email addresses (to ensure they are not misinterpreted as file references), and multi-line prompts with @file on a non-first line.

Using Gemini Code Assist

The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.

Invoking Gemini

You can request assistance from Gemini at any point by creating a comment using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands on the current page.

Feature	Command	Description
Code Review	/gemini review	Performs a code review for the current pull request in its current state.
Pull Request Summary	/gemini summary	Provides a summary of the current pull request in its current state.
Comment	@gemini-code-assist	Responds in comments when explicitly tagged, both in pull request comments and review comments.
Help	/gemini help	Displays a list of available commands.

Customization

To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with :thumbsup: and :thumbsdown: on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here.

You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.

[^1]: Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution.