feat: Add predefined suggestions after finishing the planning session

[alex-alecu]

Context

Following the example from Claude Code after a planning session:

The context management is the most important job of tools using AI models. By starting a fresh session with just the plan as the first message, we maximize the efficiency of the implementation.

It's also cheaper to start with a fresh context - the CLI will not send on every request all the context it took build the plan.

Implementation

Added plan follow-up flow in plan-followup.ts (line 25).

• Added plan follow-up flow in /Users/alex/Documents/code/kilo-code/kilo/packages/opencode/src/kilocode/plan-followup.ts:25.

  • Asks Ready to implement?
  • Handles Start new session, Continue here, and custom text
  • Injects synthetic user messages using the same pattern as plan tools
  • Publishes TuiEvent.SessionSelect and starts the new session loop for the new-session path

• Wired it into the existing loop break point in /Users/alex/Documents/code/kilo-code/kilo/packages/opencode/src/session/prompt.ts:335 with kilocode_change markers.

  • If plan agent just finished, it now calls PlanFollowup.ask(...)
  • continue re-enters the loop when needed; otherwise it exits as before

• Added tests in /Users/alex/Documents/code/kilo-code/kilo/packages/opencode/test/kilocode/plan-followup.test.ts:95.

  • Covered dismissed, continue here, custom text, start new session, empty assistant text, aborted signal, and blank custom answer

Screenshots

How to Test

Start a planning session and look for the suggestions at the end.

Get in Touch

We'd love to have a way to chat with you about your changes if necessary. If you're in the Kilo Code Discord, please share your handle here.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated (UTC)
kilo-ui-storybook	Ready	Preview, Comment	Feb 19, 2026 0:10am

[kiloconnect]

Code Review Summary

Status: No Issues Found | Recommendation: Merge

The PR introduces a plan follow-up feature that prompts users after the plan agent completes, offering options to start a new session, continue in the current session, or provide custom follow-up text. The implementation is clean and well-structured:

• Abort handling is properly implemented via Question.reject + RejectedError catch pattern, with correct cleanup of event listeners.
• Circular dependency between prompt.ts and plan-followup.ts is avoided via dynamic import() in startNew.
• Context extraction (extractContext) correctly deduplicates file reads, strips <task_result> tags, and truncates at the character limit with clean line-boundary truncation.
• Type narrowing in isTool guard properly narrows ToolPart to the completed state variant.
• Test coverage is comprehensive with 14 tests covering all code paths: dismiss, continue, custom text, new session, empty text, abort (pre and mid-question), blank answers, and all extractContext edge cases.

All existing review comments have been addressed by the author.

Files Reviewed (3 files)

• packages/opencode/src/kilocode/plan-followup.ts - New file, 204 lines. Core plan follow-up logic.
• packages/opencode/src/session/prompt.ts - 10-line integration point in the session loop.
• packages/opencode/test/kilocode/plan-followup.test.ts - New file, 555 lines. Comprehensive test suite.

[marius-kilocode]

Not really sure what this is doing? Is this user output we need to later localize?

[alex-alecu]

If we start a new session just with the plan, the coding agent will do some research first to make sure it understands the plan.

By providing a small amount of research context from the planning agent we reduce drastically the research the new coding agent needs to make.

[marius-kilocode]

What if the beginning of the research for the plan wasn't valuable? What CC actually does instead they reference the most important files that are relevant on compaction instead. I would question if the first 10k characters have always high entropy.