save-trajectory

You can export a conversation trajectory to JSON by running this skill, which captures user messages, assistant responses, and tool interactions into a structured file for analysis and model refinement.

A trajectory JSON file is a structured export of the complete conversation sequence, needed when you want to replay, audit, or benchmark agent behavior for offline analysis and fine-tuning data collection.

You can save chat sessions for debugging and auditing using this skill, which requires only the Python standard library and writes timestamped trajectory files directly to your project directory.

Yes, you can log agent interactions for fine-tuning data collection using this skill, which relies solely on the Python standard library to export the full conversation sequence into JSON format.

No, you do not need external packages to export conversation trajectories, as this skill requires no dependencies beyond the Python standard library to write the JSON files.

Exported trajectory JSON files are saved to the .evolve/trajectories directory in your project root, using timestamped filenames to automatically avoid any versioning or naming collisions.