Autoresearch
SkillMonitoring & opsYour AI can run bounded research experiments, testing hypotheses and keeping the changes that improve benchmark results. autoresearch structures this as a loop that measures benchmark evidence and records what fails. Use it when you want to iteratively improve a research metric or benchmark a hypothesis.
Available today. Use it from your connected AI after setup.
No other account needed.
After adding it, ask your AI to optimize a research metric or benchmark a specific hypothesis. It will run the experiment loop and keep what works.
Then ask your AI: use the Autoresearch skill
What your AI can do with it
- Run bounded experiment loops to optimize a research metric
- Test a research hypothesis against benchmark evidence
- Keep the changes that improve benchmark results
- Record failed experiments so the lessons are not lost
- Iterate on model, retrieval, or evaluation performance
What this skill tells your AI
The instructions your AI receives, as published by companion-inc/feynman in skills/autoresearch/SKILL.md and read by ahel’s review.
Run the /autoresearch workflow. The slash command expands the full workflow instructions in the active session; do not try to read a relative prompt-template path from the installed skill directory.
Optional tools used when visible: init_experiment, run_experiment, log_experiment. Without those tools, run the benchmark through the available shell/tooling and record benchmark result, evidence, and decision in the session files.
Session files: autoresearch.md, autoresearch.sh, autoresearch.jsonl
Signals
- GitHub stars
- 9k
- Forks
- 1k
- Last commit
- Sep 2026
Advanced
- Catalog kind
- skill
- Gateway key
autoresearch-companion-inc- Source
- github.com/companion-inc/feynman