Automatically optimize files by metrics

Optimize any file by a measurable metric in autonomous experiment loops at configurable intervals

When optimizing a single file's performance, quality, or cost over time requires hands-off autonomous loops with clear metric tracking.

• · target file (code, config, content, etc.)
• · measurable metric (test pass rate, latency, word count, engagement, etc.)
• · optimization strategy (mutation, A/B variants, etc.)

• · series of improved versions
• · metric improvement trajectory
• · loop iteration log

• · 5 slash commands (/autoresearch-start, /autoresearch-status, /autoresearch-stop, /autoresearch-eval, /autoresearch-merge)
• · 8 evaluators (metric-based, LLM, hybrid, etc.)
• · stdlib Python scheduler

• · metric is computable (not subjective)
• · loop interval is feasible (10min to monthly)
• · file has enough mutation surface area to improve

• · Metric gaming: optimization inflates metric without improving actual quality
• · Loop runs into local maxima and stops improving
• · Mutation strategy unsuited to file type (e.g., changing prompt wording won't help compiled code)

• · 5 slash commands for lifecycle management
• · 8 evaluators (metric, LLM, A/B, etc.)
• · Configurable loop intervals (10min to monthly)
• · Iteration log for audit trail