autoresearch-x

Autonomous iteration automates research tasks by coordinating exploration with evidence chains to prove outcomes. It enforces guardrail hooks and a one-change-per-iteration discipline, ensuring reliable experimentation and structured tracking for reproducible results.

Automate iterative debugging by running debug mode, which applies defined phases to isolate issues. It logs per-iteration results in results.tsv and per-commit notes in the iterations directory to maintain structured evidence chains for every change.

Yes, autonomous iteration supports code optimization and investigative data analysis. Optimize mode improves code performance, while investigate mode explores data, both tracking reproducible evidence chains in results.tsv.

Start an autonomous iteration session to draft a program.md file. This initializes the coordinated exploration process, applying guardrails and logging per-iteration results to ensure reproducible outcomes for your experiments.

Track reproducibility by using branching and team-mode workflows to scale tasks. The framework maintains evidence-chain tracking for every decision, logging per-iteration results in results.tsv and per-commit notes in the iterations directory.