feat: wire symbols/painspots into next-task + downgrade exploration-agent to sonnet

[avifenesh]

Summary

• exploration-agent: queries symbols for primary files to expose exported names and blast radius of changes
• planning-agent: queries painspots (hotspot × complexity × bug density) for planned files; uses these as high-risk signals in the implementation plan
• exploration-agent model: Opus → Sonnet (pre-fetched structured data reduces need for open-ended reasoning)

Why the model downgrade is safe

exploration-agent now receives pre-fetched structured signals from agent-analyzer:

• Hotspots, bugspots, coupling, ownership (already had these)
• Symbols with exported names per primary file (new: reveals blast radius)

The agent curates and formats this structured data. Complex reasoning (dependency strategy, risk assessment) happens in planning-agent (Opus), which receives the enriched report. Same pattern that worked for onboard-agent and can-i-help-agent.

Test plan

• Run /next-task on a real GitHub issue - verify exploration report includes symbol exports section
• Verify exploration-agent uses Sonnet (check modelUsage in output JSON)
• Verify planning report includes painspots risk table when repo-intel.json exists
• Verify planning-agent still uses Opus

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.
To continue using code reviews, add credits to your account and enable them for code reviews in your settings.

[gemini-code-assist]

Summary of Changes

Hello, 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 enhances the risk assessment capabilities of the exploration and planning agents by integrating symbol information and painspot analysis. It also optimizes the exploration agent's performance by downgrading its model to Sonnet, which is now feasible due to pre-fetched structured signals. These changes aim to provide more informed and efficient task planning and code review processes.

Highlights

• Exploration Agent Update: The exploration agent now uses symbols to expose exported names and the blast radius of changes, enhancing risk assessment.
• Planning Agent Enhancement: The planning agent now incorporates painspots (hotspot × complexity × bug density) for planned files, using these as high-risk signals in the implementation plan.
• Model Downgrade: The exploration agent's model has been downgraded from Opus to Sonnet, leveraging pre-fetched structured data to reduce the need for open-ended reasoning.

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.

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-code-assist]

Code Review

This pull request wires up symbols and painspots data into the exploration and planning agents, respectively, to provide them with more context about code structure and risk. It also refactors several files to use a centralized function for locating the state directory, which is a great improvement for maintainability. The exploration agent's model is also downgraded to Sonnet, with the justification that more structured data reduces the need for a more powerful model. My feedback focuses on improving error handling and code style for better debuggability and consistency.