Iteratively modify, execute, and evaluate targets against a defined metric.
npx skills add https://github.com/estebanrfp/gos --skill autoresearch-estebanrfp
Please help me install this Agent Skill. Skill: Autoresearch Source: https://github.com/estebanrfp/gos/tree/main/skills/autoresearch Command: npx skills add https://github.com/estebanrfp/gos --skill autoresearch-estebanrfp
This Skill removes the manual grind from iterative optimization by letting an agent baseline, test, measure, and refine a target until the best-performing change is found.
Ask the agent to optimize a measurable target, define the metric to track, and let it run baseline and improvement experiments until it reports the best result.
High-intent search queries and answers about installing and using this skill.
Automated iterative optimization modifies a measurable target, executes it, and evaluates the results against a baseline metric to keep or revert changes. This automates prompt tuning and code performance benchmarking by establishing a reference run and evaluating modifications iteratively.
Yes, overnight experimentation runs can be scheduled using cron-based resumption. The agent executes long-duration metric sweeps and logs keep-or-revert decisions for every iteration, allowing benchmarking and configuration sweeps to improve overnight without manual intervention.
Baseline-first experimentation records a reference result before any modifications are made. After applying a controlled change, the agent executes a test, extracts the target metric from the command output, and compares it against that baseline to measure true improvements.
Yes, automated configuration sweeps are workspace-safe. All experimental modifications are confined strictly to the agent workspace, and the system logs all keep-or-revert decisions so you can review outcomes without risking uncontrolled changes to your broader environment.
The best way to set up metric extraction is to define a quantifiable target metric in your command output before starting. The optimization loop then parses this metric after each execution to decide whether the applied changes should be kept or reverted.
usecompaiExecute an autonomous optimize-measure-keep/discard loop on target files.
Veloxia-agencyOptimize a target file by iteratively editing and evaluating a measurable metric.
kausthubh-coderSchedule and execute recurring background tasks in Convex applications using cron expressions.
phoenicisfGenerate, validate, and refine SKILL.md definitions for Claude skills.
