Automate iterative experiments to optimize a measurable metric.
npx skills add https://github.com/maragudk/fabrik --skill autoresearch-maragudk
Please help me install this Agent Skill. Skill: autoresearch Source: https://github.com/maragudk/fabrik/tree/main/skills/autoresearch Command: npx skills add https://github.com/maragudk/fabrik --skill autoresearch-maragudk
Automates autonomous experimentation to optimize a measurable metric by iteratively branching, measuring, and deciding which changes to keep.
Set a goal and a verify command, then start the autoresearch loop.
High-intent search queries and answers about installing and using this skill.
Automated metric optimization works by iteratively branching, measuring, and deciding which changes to keep. It applies to any domain with a quantifiable objective, such as code performance or build size, by running autonomous experiments until the goal is met.
To start autonomous experimentation, you need a clear numeric goal, a shell command to verify the metric, an optional guard to block unsafe changes, and an optional scope to limit file modifications. The process operates in two phases: plan and loop.
Yes, you can restrict file modifications by defining an optional scope. This limits the autonomous experimentation to specific areas, preventing unintended changes to the rest of the codebase during the optimization loop.
Prevent unsafe code changes by defining an optional guard before starting the loop. This guard acts as a safety check to block modifications that violate your constraints while the system attempts to optimize the measurable metric.
Yes, automated metric optimization works for reducing build size. It supports any task with a numeric objective and a shell command to verify the outcome, iterating through branching and measurement to achieve the target size.
datamonsterrAutomate iterative code experiments with git-based logging and metric tracking.
guccangCreate failing tests before implementing minimal production code.
jh941213Modify, validate, and selectively keep or discard code changes in iterative experiments.
fcouryAnalyze Rust libraries for runtime performance and memory optimization opportunities.
