Skip to content

feat(core): experimental in-progress steering hints (1 of 3)#19008

Merged
joshualitt merged 2 commits intomainfrom
jl/model-steering-1
Feb 17, 2026
Merged

feat(core): experimental in-progress steering hints (1 of 3)#19008
joshualitt merged 2 commits intomainfrom
jl/model-steering-1

Conversation

@joshualitt
Copy link
Contributor

@joshualitt joshualitt commented Feb 13, 2026 •
edited
Loading

Roll-up PR here:
#18973

This is a rebase / refactor of:
#18783

Part of a series to address issue #18782

This change lands a few supplementary pieces for testing as well as some utility functions

@gemini-code-assist
Copy link
Contributor

gemini-code-assist bot commented Feb 13, 2026

Summary of Changes

Hello @joshualitt, I'm Gemini Code Assist1! 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 introduces foundational elements for experimental in-progress steering hints, primarily by adding a robust UI testing framework (AppRig) and a specialized flash-lite-helper model. These changes facilitate the development and testing of features that allow users to guide the AI model more effectively during its operation, with the flashLiteHelper specifically designed to generate concise acknowledgements for these steering inputs. Additionally, core configuration initialization has been made more resilient to multiple calls.

Highlights

  • New UI Testing Utility (AppRig): Introduced AppRig, a comprehensive utility for testing the CLI's user interface. This allows for simulating user interactions, managing configuration, mocking shell commands, and asserting UI output in an isolated test environment.
  • Experimental In-Progress Steering Hints: Added a new flash-lite-helper model configuration and a flashLiteHelper utility. This enables the system to generate short, friendly acknowledgements for user steering updates during in-progress tasks, enhancing the interactive experience.
  • Improved Configuration Initialization: Refactored the core Config class initialization logic to deduplicate multiple calls, ensuring that setup processes like storage and tool discovery are only performed once, improving robustness and efficiency.
  • Enhanced Evaluation Test Setup: Updated the Vitest configuration for evaluation tests and introduced appEvalTest to streamline behavioral evaluations using the new AppRig testing framework.
Changelog
  • docs/get-started/configuration.md
    • Added flash-lite-helper model configuration to the documentation.
  • evals/app-test-helper.ts
    • Added appEvalTest function to provide a helper for running behavioral evaluations using AppRig.
  • evals/vitest.config.ts
    • Updated Vitest configuration to include resolve conditions, node environment, globals, react alias, setupFiles, and server.deps.inline for evaluation tests.
  • packages/cli/src/test-utils/AppRig.test.tsx
    • Added tests for the new AppRig class, verifying its ability to render the app and handle messages.
  • packages/cli/src/test-utils/AppRig.tsx
    • Added the AppRig class, a new testing utility for rendering the CLI app, managing configuration, mocking services, and simulating user interactions.
  • packages/cli/src/test-utils/MockShellExecutionService.ts
    • Added MockShellExecutionService to provide a mock implementation for shell command execution in tests.
  • packages/cli/src/test-utils/fixtures/simple.responses
    • Added a JSON fixture for mock model responses in testing.
  • packages/cli/src/test-utils/render.tsx
    • Imported SessionStatsProvider and wrapped the AppContainer with it in the test renderer.
  • packages/cli/src/ui/hooks/useGeminiStream.test.tsx
    • Updated the mock for useSessionStats to use importOriginal for better mock isolation.
  • packages/core/src/config/config.test.ts
    • Modified the initialize test to verify deduplication of multiple calls to config.initialize() instead of throwing an error.
  • packages/core/src/config/config.ts
    • Replaced the initialized boolean with an initPromise to ensure Config initialization is deduplicated and handled gracefully across multiple calls.
  • packages/core/src/config/defaultModelConfigs.ts
    • Added a new flash-lite-helper model configuration with specific temperature, maxOutputTokens, and thinkingBudget settings.
  • packages/core/src/index.ts
    • Exported baseLlmClient.js and flashLiteHelper.js from the core package.
  • packages/core/src/services/test-data/resolved-aliases-retry.golden.json
    • Added flash-lite-helper model configuration to the golden test data for resolved aliases with retry.
  • packages/core/src/services/test-data/resolved-aliases.golden.json
    • Added flash-lite-helper model configuration to the golden test data for resolved aliases.
  • packages/core/src/utils/flashLiteHelper.test.ts
    • Added tests for flashLiteHelper functions, covering truncation, text generation, and steering acknowledgement messages.
  • packages/core/src/utils/flashLiteHelper.ts
    • Added flashLiteHelper utility functions, including generateFlashLiteText, truncateFlashLiteInput, and generateSteeringAckMessage, along with related constants and instructions for user steering.
  • schemas/settings.schema.json
    • Updated the modelConfigs schema to include the new flash-lite-helper alias in its default configuration.
Activity
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 👍 and 👎 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.

Footnotes

  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. ↩

@gemini-cli gemini-cli bot added area/agent Issues related to Core Agent, Tools, Memory, Sub-Agents, Hooks, Agent Quality 🔒 maintainer only ⛔ Do not contribute. Internal roadmap item. labels Feb 13, 2026
gemini-code-assist[bot]
gemini-code-assist bot reviewed Feb 13, 2026
View reviewed changes
Copy link
Contributor

@gemini-code-assist gemini-code-assist bot left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Code Review

The pull request introduces experimental steering hints using a lightweight model (gemini-2.5-flash-lite), infrastructure for in-process application testing (AppRig), and a refactor of core configuration initialization. A high-severity prompt injection vulnerability has been identified in flashLiteHelper.ts due to unsafe concatenation of user-provided steering hints into LLM prompts without proper escaping or sanitization, which is critical given the agent's access to powerful tools like shell execution. Implementing robust escaping for delimiters or using structured data formats for prompt construction is recommended to mitigate this risk. Additionally, several instances of string truncation logic violate the grapheme-aware rule for multi-byte Unicode characters.

github-advanced-security[bot]
github-advanced-security bot found potential problems Feb 13, 2026
View reviewed changes
@github-actions
Copy link

github-actions bot commented Feb 13, 2026 •
edited
Loading

Size Change: +6.55 kB (+0.03%)

Total Size: 24.5 MB

Filename Size Change
./bundle/gemini.js 24.5 MB +6.55 kB (+0.03%)
ℹ️ View Unchanged
Filename Size
./bundle/sandbox-macos-permissive-open.sb 890 B
./bundle/sandbox-macos-permissive-proxied.sb 1.31 kB
./bundle/sandbox-macos-restrictive-open.sb 3.36 kB
./bundle/sandbox-macos-restrictive-proxied.sb 3.56 kB
./bundle/sandbox-macos-strict-open.sb 4.82 kB
./bundle/sandbox-macos-strict-proxied.sb 5.02 kB

compressed-size-action

@joshualitt joshualitt force-pushed the jl/model-steering-1 branch 3 times, most recently from c1a5fd3 to 3e0f67a Compare February 13, 2026 19:02
@joshualitt joshualitt marked this pull request as ready for review February 13, 2026 19:30
@joshualitt joshualitt requested review from a team as code owners February 13, 2026 19:30
gundermanc
gundermanc reviewed Feb 17, 2026
View reviewed changes
gundermanc
gundermanc reviewed Feb 17, 2026
View reviewed changes
@joshualitt
feat(core): experimental in-progress steering hints (1 of 3)
ac1e154 
Roll-up PR here:
#18973

This is a rebase / refactor of:
#18783

Part of a series to address issue #18782
@joshualitt joshualitt added this pull request to the merge queue Feb 17, 2026
Merged via the queue into main with commit 55c628e Feb 17, 2026
26 of 27 checks passed
@joshualitt joshualitt deleted the jl/model-steering-1 branch February 17, 2026 23:09
@gemini-code-assist gemini-code-assist bot mentioned this pull request Feb 18, 2026
Merged
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

@gundermanc gundermanc gundermanc approved these changes

@mattKorwel mattKorwel Awaiting requested review from mattKorwel

+1 more reviewer

@gemini-code-assist gemini-code-assist[bot] gemini-code-assist[bot] left review comments

Reviewers whose approvals may not affect merge requirements

Assignees

No one assigned

Labels

area/agent Issues related to Core Agent, Tools, Memory, Sub-Agents, Hooks, Agent Quality 🔒 maintainer only ⛔ Do not contribute. Internal roadmap item.

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

2 participants

@joshualitt @gundermanc