Research Loop

Run autonomous, metric-driven experiments on a version-controlled implementation against a fixed trusted evaluation harness. Use when the user asks to improve eval pass rate, benchmark score, prompt or policy quality, performance, cost, or another measurable outcome through repeated hypothesis, change, evaluate, and keep-or-discard cycles. Protect generalization with holdout gates, immutable evidence provenance, and replay-safe recovery. Reject hardcoded cases, benchmark leakage, scenario-specific patches, and harness gaming. Do not use for one-off debugging, ordinary test fixing, or Agent Skill artifact optimization.

Emlembow Updated
Emlembow Skills: portable Agent Skills distributed through npx skills to Codex, Claude Code, and other compatible agents

File contents

Emlembow/skills commit 83362f1632

Frequently asked questions

npx get-skill add emlembow/research-loop