Run autonomous research experiments
autoresearchskillsetup L1★0
Sheshiyer/skill-clusters ↗What it does
Run iterative keep-or-discard experiment loops
Best for
When running overnight experiment loops with strict keep-or-discard decisions on single-variable changes.
Inputs
- · baseline metric
- · experiment variables
- · decision rules
Outputs
- · experiment log
- · winning changes
- · discarded ideas
- · next batch recommendations
Requires
- · autoresearch repo or business-loop workspace
Preconditions
Defined decision metric before running; bounded batch size (5-20 unless explicitly open-ended)
Failure modes
Ambiguous results cause reversion; baseline not logged makes comparison impossible
Trust signals
- · One-variable-per-experiment discipline
- · reversion log on failure
- · baseline-first approach