loumaciel
Joined 4 years ago
Karma
38
Submissions
14
Member since
Mar 21
Joined 4 years ago
It saves the payloads into SQLite, maps them, and exposes tools for the model to run python against them. Works very well.
The benchmark harness and datasets are in the repo if anyone wants to reproduce or extend the tests. Curious if others have run into the same context compaction issues with tool-heavy agents.