Execute metric-driven optimization loops using a spec file and scoring gates.
npx skills add https://github.com/FFCfelps1/SGPPF_Maua --skill ce-optimize-ffcfelps1
Please help me install this Agent Skill. Skill: ce-optimize Source: https://github.com/FFCfelps1/SGPPF_Maua/tree/main/.gemini/skills/ce-optimize Command: npx skills add https://github.com/FFCfelps1/SGPPF_Maua --skill ce-optimize-ffcfelps1
This Skill addresses the challenge of optimizing complex metrics, like clustering quality, search relevance, or build performance, through iterative, metric-driven loops that converge on the best solution.
Run the ce-optimize skill with an optimization spec file, e.g., 'ce-optimize references/example-hard-spec.yaml'.
High-intent search queries and answers about installing and using this skill.
You can optimize clustering quality or search relevance by running metric-driven iterative loops that test parallel experiments, measure results against hard gates or LLM-as-judge scores, and keep improvements to converge on the best solution.
Metric-driven optimization works by executing iterative loops using a defined spec file, applying hard gates and LLM-as-judge scores to systematically evaluate parallel experiments and converge on the best solution for complex metrics.
The best way to set up an iterative optimization loop is defining a spec file that outlines your optimization goal, measurement harness, and scoring rubric, then running the optimization process to systematically select the best model parameters.
Yes, you need a spec file to run metric-driven optimization loops. The spec file must define the optimization goal, measurement harness, and scoring rubric to properly execute the iterative experiments and evaluate complex metrics.
Yes, you can use LLM-as-judge scores to evaluate complex metrics like build performance. The optimization process applies these scores alongside hard gates to measure parallel experiments and converge on the best solution.
You should use an iterative optimization loop instead of manual tuning when optimizing complex metrics like clustering quality or search relevance that require systematic experimentation, parallel testing, and measurable convergence via hard gates or LLM-as-judge scores.
SantiagoisperRun metric-driven iterative optimization loops with parallel experiments and LLM-as-judge scores.
shsteimerSearch AEM Live documentation and return ranked results with relevance scores.
DieStokAutomate metric-driven optimization loops with controlled experiments and LLM judging.
ovachieverImplement, evaluate, and optimize scikit-learn machine learning models with pipelines.
