MoonshotAI / MoonshotAI/kimi-code

/fork-session — mid-session snapshot for safe model experimentation

Open
#3,389 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

Dominant language
TypeScript
Stars
7.5k
Forks
1.2k
Avg merge
11h 53m
Merged PRs (30d)
350

Description

When I am mid-session and run into peak hours, or my credits are exhausted on a premium model and I have to wait, switching to a less smart model will pollute the session irreversibly and is prone to compactions.

A /fork-session command should create a duplicate of the currently loaded session. The previous session stays untouched. This lets me experiment with the session using different models while preserving the state of the original session.

Key desired behavior:

  • Triggerable from inside an active session (slash command).
  • Creates a copy of the session state (transcript, context, attached tools/skills) at the fork point.
  • The fork is fully independent — changes in the fork do not bleed back into the original.
  • The original session remains available exactly as it was, no compaction or rewriting forced on it.
  • Allows trying a cheaper / different model on the fork without risking irreversible degradation of the premium-model run.

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start by tracing how active-session slash commands access the currently loaded session and how transcript, context, and attached tools or skills are represented. Define the fork point and verify that the original remains unchanged while the fork can evolve independently and use a different model. Done means /fork-session works inside an active session and preserves all stated isolation guarantees.

Written by the indexing model from the issue text.

Assessment

Tech stack
typescript
Domain
cli
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Active
Clarity
Mostly clear
Newbie friendliness
35/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.