GenPark Agent State Snapshot & Time-Travel Skill

Local MCP server that snapshots agent conversation state and rolls back to earlier points, pruning divergent hallucination paths.

What you can do

• Snapshot agent conversation state at chosen points.
• Roll back to an earlier snapshot (time travel) to discard divergent paths.
• Prune branches that lead to hallucinated or off-track results.
• Run as a local MCP server or import it as a Python library.

When to use it

Useful when an assistant's long-running conversation or agent workflow drifts off track and you want to return to a known-good state instead of restarting. Suited to experimentation with agent DAGs and multi-step reasoning where divergent branches need discarding.

Requirements

Runs as a local process on the user's machine (desktop only), started with Python; the README states it uses only the Python standard library, so no third-party runtime dependencies. No authentication, environment variables or headers are declared.