Adds a test case to `settings.test.ts` to specifically verify
the correct resolution of multiple environment variables concatenated
within a single string value (e.g., ${HOST}:${PORT} ).
Refactors the `resolveEnvVarsInObject` function in settings to
explicitly handle primitive types (null, undefined, boolean, number)
at the beginning of the function. This clarifies the logic for
subsequent string, array, and object processing.
This commit introduces the ability to use system environment variables
within the settings files (e.g., `settings.json`). Users can now
reference environment variables using the `${VAR_NAME}` syntax.
This enhancement improves security and flexibility, particularly
for configurations like MCP server settings, which often require
sensitive tokens.
Previously, to configure an MCP server, a token might be directly
embedded:
```json
"mcpServers": {
"github": {
"env": {
"GITHUB_PERSONAL_ACCESS_TOKEN": "pat_abc123"
}
// ...
}
}
```
With this change, the same configuration can securely reference an
environment variable:
```json
"mcpServers": {
"github": {
"env": {
"GITHUB_PERSONAL_ACCESS_TOKEN": "${GITHUB_PERSONAL_ACCESS_TOKEN}"
}
// ...
}
}
```
This allows users to avoid storing secrets directly in configuration files.
Address multiple possible memory leaks found bystatic analysis of the codebase. The primary source of the leaks was event listeners on child processes and global objects that were not being properly removed, potentially causing their closures to be retained in memory indefinitely particularly for processes that did not exit.
There are two commits. A larger one made by gemini CLI and a smaller one by me to make sure we always disconnect child processes as part of the cleanup methods. These changes may not actually fix any leaks but do look like reasonable defensive coding to avoid leaking event listeners or child processes.
The following files were fixed:
This is Gemini's somewhat overconfident description of what it did.
packages/core/src/tools/shell.ts: Fixed a leak where an abortSignal listener was not being reliably removed.
packages/cli/src/utils/readStdin.ts: Fixed a significant leak where listeners on process.stdin were never removed.
packages/cli/src/utils/sandbox.ts: Fixed leaks in the imageExists and pullImage helper functions where listeners on spawned child processes were not being removed.
packages/core/src/tools/grep.ts: Fixed three separate leaks in the isCommandAvailable check and the git grep and system grep strategies due to un-removed listeners on child processes.
packages/core/src/tools/tool-registry.ts: Corrected a leak in the execute method of the DiscoveredTool class where listeners on the spawned tool process were not being removed.
# Add .gitignore-Aware File Filtering to gemini-cli
This pull request introduces .gitignore-based file filtering to the gemini-cli, ensuring that git-ignored files are automatically excluded from file-related operations and suggestions throughout the CLI. The update enhances usability, reduces noise from build artifacts and dependencies, and provides new configuration options for fine-tuning file discovery.
Key Improvements
.gitignore File Filtering
All @ (at) commands, file completions, and core discovery tools now honor .gitignore patterns by default.
Git-ignored files (such as node_modules/, dist/, .env, and .git) are excluded from results unless explicitly overridden.
The behavior can be customized via a new fileFiltering section in settings.json, including options for:
Turning .gitignore respect on/off.
Adding custom ignore patterns.
Allowing or excluding build artifacts.
Configuration & Documentation Updates
settings.json schema extended with fileFiltering options.
Documentation updated to explain new filtering controls and usage patterns.
Testing
New and updated integration/unit tests for file filtering logic, configuration merging, and edge cases.
Test coverage ensures .gitignore filtering works as intended across different workflows.
Internal Refactoring
Core file discovery logic refactored for maintainability and extensibility.
Underlying tools (ls, glob, read-many-files) now support git-aware filtering out of the box.
Co-authored-by: N. Taylor Mullen <ntaylormullen@google.com>
- Introduces a suite of tests for the hook, covering various scenarios including:
- Successful tool execution
- Tool not found errors
- Errors during
- Errors during tool execution
- Tool confirmation (approved and cancelled) - (currently skipped)
- Live output updates - (currently skipped)
- Cancellation of tool calls (before execution and during approval) - (currently skipped)
- Execution of multiple tool calls
- Preventing scheduling while other calls are running - (currently skipped)
- Includes tests for the utility function to ensure correct mapping of tool call states to display objects.
- Mocks dependencies like , , and individual instances.
- Uses fake timers to control asynchronous operations.
Note: Some tests involving complex asynchronous interactions (confirmations, live output, cancellations) are currently skipped due to challenges in reliably testing these scenarios with the current setup. These will be addressed in future work.
This change introduces a small delay after the first Ctrl+C press, prompting the user to press Ctrl+C again to exit. This helps prevent accidental termination of the application.
- Added `exitOnCtrlC={false}` to the Ink render options in `gemini.tsx` to enable custom Ctrl+C handling.
- Implemented logic in `App.tsx` to:
- Display "Press Ctrl+C again to exit." for 2 seconds after the first Ctrl+C.
- Exit the application if Ctrl+C is pressed again during this period.
- Revert to normal operation if the second Ctrl+C is not pressed within the timeout.
- Defined a constant `CTRL_C_PROMPT_DURATION_MS` for the timeout duration.
- Consolidates consecutive model messages into a single message in the chat history.
- This prevents multiple model messages from being displayed in a row, improving readability.
- This may also address some instances of 500 errors that could have been caused by multiple, rapidly succeeding model messages.
- Adds tests for the new consolidation logic.
Fixes https://b.corp.google.com/issues/421010429
- Moves MCP tool discovery logic from ToolRegistry into a new, dedicated MCP client (mcp-client.ts and mcp-tool.ts).
- Updates ToolRegistry to utilize the new MCP client.
- Adds comprehensive tests for the new MCP client and its integration with ToolRegistry.
Part of https://github.com/google-gemini/gemini-cli/issues/577
This change detects the most recent git commit short hash and writes it to the `GIT_COMMIT_INFO` constant in `packages/cli/src/generated/git-commit.sh`, optionally appending the string "(local modifications)" if additional local changes after that commit are detected.
If set, this string is displayed in the `/about` dialog as well as passed into the `/bug` template.
Example:
```
> /about
╭───────────────────────────────────────────────────────────────────────────╮
│ │
│ About Gemini CLI │
│ │
│ CLI Version development │
│ Git Commit 43370ab (local modifications) │
│ Model gemini-2.5-pro-preview-05-06 │
│ Sandbox sandbox-exec (minimal) │
│ OS darwin v23.11.0 │
│ │
╰───────────────────────────────────────────────────────────────────────────╯
```
Additionally, this change updates `.gitignore` to ignore the generated files, `scripts/clean.sh` to remove them, and adds a `npm run generate` stage for this and any other generators we need to write.
- Implements cancellation for Gemini requests while they are actively being processed by the model.
- Extends cancellation support to the logic within tools. This allows users to cancel operations during the phase where the system is determining if a tool execution requires user confirmation, which can include potentially long-running pre-flight checks or LLM-based corrections.
- Underlying LLM calls for edit corrections (within and ) and next speaker checks can now also be cancelled.
- Previously, cancellation of the main request was not possible until text started streaming, and pre-execution checks were not cancellable.
- This change leverages the updated SDK's ability to accept an abort token and threads s throughout the request, tool execution, and pre-execution check lifecycle.
Fixes https://github.com/google-gemini/gemini-cli/issues/531
- Removes the sqlite3 dependency and refactors the logging mechanism to use a JSON file (logs.json) instead of a database.
- This change is a temporary workaround to address issues with macOS sandboxing that were caused by the SQLite native module.
- Storing all logs in a single JSON file may introduce scalability concerns in the future.
Fixes https://github.com/google-gemini/gemini-cli/issues/522
- Enhance `checkNextSpeaker` to handle cases where the last message is a function response or an empty model message.
- If the last message is a function response, the model should speak next.
- If the last message is an empty model message, the model should speak next.
- This ensures more robust and accurate determination of the next speaker in the conversation, completing the fix for the issue.
- Updated tests.
Fixes https://github.com/google-gemini/gemini-cli/issues/551
- Corrects an issue where the `nextSpeakerCheck` would determine the model should speak next, but the models response was not properly propagated due to a missing `yield*` in a recursive call within `sendMessageStream`.
- This change ensures that when the model is designated as the next speaker, its generated content is correctly unwoven and returned, allowing the conversation to proceed as expected.
Part of https://github.com/google-gemini/gemini-cli/issues/551
- Addresses a Gemini model bug where it may return an empty content object after a function response.
- Previously, the SDK attempted to inject an empty model message, which could disrupt curated history.
- This change modifies our custom class to detect this scenario using an utility and avoid pushing an unnecessary empty model message, thus preserving history integrity.
Workaround for https://b.corp.google.com/issues/420354090
Part of https://github.com/google-gemini/gemini-cli/issues/551
- Copied the `Chat` class from `@google/genai` into `packages/server/src/core/geminiChat.ts`.
- This change is in preparation for future modifications to the chat handling logic.
- Updated relevant files to use the new `GeminiChat` class.
Part of https://github.com/google-gemini/gemini-cli/issues/551
- Upgraded the @google/genai SDK from ^0.13.0 to ^1.0.1.
- Addressed a related type error in `packages/server/src/tools/edit.test.ts` by updating a type assertion.
Part of https://github.com/google-gemini/gemini-cli/issues/551
- Updated the diff generation in `edit.ts` and `write-file.ts` to include the `ignoreWhitespace: true` option.
- This ensures that whitespace-only changes are not highlighted in the diff output when there are other content modifications, making the diffs cleaner and easier to review.
- Extract default diffing options into single source of truth.
Fixes https://github.com/google-gemini/gemini-cli/issues/548
- Introduced a 'validating' state for tool calls to prevent the input box from reappearing while waiting for a tool's `shouldConfirmExecute` method to complete.
- When a tool call is initiated, it's now immediately set to a 'validating' status. This ensures the UI remains in a busy/responding state.
- `useGeminiStream` now considers the 'validating' state as part of `StreamingState.Responding`.
- `useToolScheduler` has been updated to:
- Set the initial status of new tool calls to 'validating'.
- Asynchronously perform the `shouldConfirmExecute` check.
- Transition to 'awaiting_approval' or 'scheduled' based on the check's outcome.
- This resolves an issue where a slow `shouldConfirmExecute` could lead to the input prompt becoming active again before the tool call lifecycle was fully determined. While 'validating' is currently treated similarly to 'executing' in the UI, this new state provides a foundation for more customized user experiences during this phase in the future.
Fixes https://github.com/google-gemini/gemini-cli/issues/527
- Enhanced by:
- Mocking and utilities (, ) to allow for more focused unit testing of .
- Adding comprehensive tests for the private method, covering new and existing file scenarios, as well as error handling.
- Expanding tests for and to verify behavior with the new content correction logic, including diff generation and directory creation.
- Refined by:
- Introducing robust mocking for and its methods (, , ) to simulate LLM interactions accurately.
- Adding extensive test scenarios for , categorized by how matches and how is processed, including direct matches, unescaping, and LLM-based corrections.
- Including tests for edge cases like no matches or multiple matches.
- Adding a utility for better test isolation.
Final fix for https://github.com/google-gemini/gemini-cli/issues/484
- Leveraged existing edit correction technology from `edit.ts` to address over-escaping issues in `write-file.ts`.
- Introduced `ensureCorrectFileContent` for correcting content in new files, where a simple "replace" isnt applicable. This uses a new LLM prompt tailored for correcting potentially problematic string escaping.
- Added caching for `ensureCorrectFileContent` to optimize performance.
- Refactored `write-file.ts` to integrate these corrections, improving the reliability of file content generation and modification.
Part of https://github.com/google-gemini/gemini-cli/issues/484
- Edit corretion leans on LLM-isms to ensure we properly fix poorly escaped content. Beacues of this we need to ensure that we don't re-run edit correction in many cases.
- To ensure this an `LruCache` has been added to capture intermediate steps of edit correction to avoid re-computations.
- Max cache size is 50 currently. This means a user can have a muti-confirmation flow of 25 items without recomputing anything (assuming they all break edit correction).
- Laid some groundwork for future testing.
Part of https://github.com/google-gemini/gemini-cli/issues/484
- Prior to this when a user would turn on auto-accept for edits we'd stop ensuring correct edits. This would result in a lot of back and forth by the model. This change also incoporates ensure correct edit into the normal execution flow.
- Added edit tests for this.
Part of https://github.com/google-gemini/gemini-cli/issues/484
- Since we're now LLM correcting a lot of problematic edits we need to also potentially trim bad edits (llms have a bad habbit of adding whitespace places).
Part of https://github.com/google-gemini/gemini-cli/issues/484
- When correcting edits prior to this if we found more than one occurrence we would try to auto-correct the old/new strings. There's no need in this situation because the tool has already provided too vague of an old_string to act upon. Instantly return.
Part of https://github.com/google-gemini/gemini-cli/issues/484
- Previously, `new_string` was assumed to be over-escaped if `old_string` was.
- This change introduces an explicit check (`newStringPotentiallyEscaped`) to determine if `new_string` itself needs correction.
- If `new_string` is potentially escaped, its corrected using an LLM call; otherwise, the original `new_string` is used.
- This avoids unnecessary corrections to `new_string` when only `old_string` was problematic.
Part of https://github.com/google-gemini/gemini-cli/issues/484
- Implement a heuristic to detect and unescape `new_string` if it appears Gemini has over-escaped it, while `old_string` is correctly formatted.
- This improves the reliability of the replace tool when the model generates an incorrectly escaped replacement string.
Part of https://github.com/google-gemini/gemini-cli/issues/484
- Refactor `ensureCorrectEdit` to clarify the correction flow for `old_string` and `new_string`.
- Only correct `new_string` if it was potentially escaped; otherwise, use the original.
- Introduce `CorrectedEditParams` and `CorrectedEditResult` interfaces for better type definition.
- Relocate `countOccurrences` for better logical grouping.
Part of https://github.com/google-gemini/gemini-cli/issues/484
Increases the threshold for rendering diff separators in the CLI's diff display. Previously, a separator was shown for gaps of more than one context line, leading to excessive separators in diffs with many small changes close together (Issue #534).
By increasing `MAX_CONTEXT_LINES_WITHOUT_GAP` to 5, we allow for more context lines before a separator is added, significantly reducing visual clutter in such diffs.
Added a test case to `DiffRenderer.test.tsx` to verify that separators are not rendered for small gaps within the new threshold.
No intentional different behavior aside for tweaks suggested from the code review of #506 Refactor: Extract console message logic to custom hook
This commit refactors the console message handling from App.tsx into a new custom hook useConsoleMessages.
This change improves the testability of the console message logic and declutters the main App component.
Created useConsoleMessages.ts to encapsulate console message state and update logic.
Updated App.tsx to utilize the new useConsoleMessages hook.
Added unit tests for useConsoleMessages.ts to ensure its functionality.
I deleted and started over on LoadingIndicator.test.tsx as I spent way too much time trying to fix it before just regenerating the tests as the code was easier to write tests for from scratch and the existing tests were not that good (I added them in the previous pull request).
Adds the following conventional readline-like keybindings:
- `Ctrl+H`: Delete the previous character.
- `Ctrl+D`: Delete the next character.
Additionally, remaps the Debug Console command from Ctrl+D to Ctrl+O, which had been first introduced in PR #486.
- The streaming state logic in `useGeminiStream.ts` has been updated.
- Previously, the loading indicator was displayed even when the system was
waiting for user confirmation on a tool call.
- This change introduces a `WaitingForConfirmation` state to ensure the
loading indicator is hidden during these confirmation prompts, improving
the user experience.
Implements recursive glob-based file search for both suggestions and execution of the `@` command.
- When typing `@filename`, suggestions will now include files matching `filename` in nested directories.
- Suggestions are sorted by path depth (shallowest first), then directories before files, then alphabetically.
- The maximum recursion depth for suggestions is set to 10.
- When executing an `@filename` command, if the file is not found directly, a recursive search (using the glob tool) is performed to locate the file.
This addresses the first request in issue #461 by allowing users to quickly reference deeply nested files without typing the full path. Also addresses b/416292478.
Tommaso Sciortino
Seth Troisi
Tolik Malibroda
jerop
Abhi
Olcan
Bryan Morgan
Mark McDonald
Eddie Santos
matt korwel
anj-s
N. Taylor Mullen
Scott Densmore
Leo
Keith Ballinger
Jacob Richman
cperry-goog
DeWitt Clinton
Miguel Solorio
Brandon Keiji
Marat Boshernitsan
Cindy Xing
Allen Hutchison
Miguel Solorio
Daniel Young Lee
sasha-gitg
cornmander
Jordan Demeulenaere