Agent transcript corpora as inspectable evidence with exact provenance
Privacy-first toolkit to harvest, normalize, and split Cursor agent transcripts into prompt-tuning datasets, skills, rules & gold exemplars for LLM agents.
Data behind Penkov, Peng and Fu: measured Co/C curves, the expert's reference fits, every fit job, the agent session transcripts, and the scripts and figures of the paper.
Browse what your coding agents wrote, and send the drafts worth keeping into your own memory. A Mnemosyne OS cartridge that reads Claude Code and Antigravity session transcripts locally. Beta.
To associate your repository with the agent-transcripts topic, visit your repo's landing page and select "manage topics."