Develop, refine, and validate Agent Skills through an eight-phase methodology.
npx skills add https://github.com/OKHP3/skillz --skill okhp3-skill-foundry
Please help me install this Agent Skill. Skill: okhp3-skill-foundry Source: https://github.com/OKHP3/skillz/tree/main/universal/okhp3-skill-foundry Command: npx skills add https://github.com/OKHP3/skillz --skill okhp3-skill-foundry
This Skill provides a comprehensive eight-phase methodology for creating and honing production-quality Agent Skills. It helps to develop clear intentions, define scopes, implement detailed design, execute evaluations, and optimize descriptions for triggering accuracy.
To use the okhp3-skill-foundry, first identify the skill you wish to create or improve. Follow the outlined phases: Architecture, Draft, Evaluation Design, Live Execution, Grading, Benchmark, Fix Loop, and Description Optimization.
High-intent search queries and answers about installing and using this skill.
A structured methodology for developing agent skills provides an eight-phase process covering architecture, drafting, evaluation design, live execution, and grading to ensure production quality.
You benchmark agent skill performance by executing live evaluations, grading results with evidence, and computing deltas to compare baseline behavior against skill-enhanced outputs.
You iterate on an agent skill by entering a fix loop after grading, using computed deltas and evidence from live execution to refine the skill implementation and description.
The best way to optimize agent skill descriptions for triggering accuracy is to apply description optimization after benchmarking, refining text based on live execution deltas and evaluation results.
Yes, this skill development methodology supports existing skill refinement by applying the fix loop, benchmarking, and description optimization phases to improve and validate current implementations.
proyecto26Structure ambiguous system design prompts into functional requirements, non-functional constraints, and out-of-scope boundaries.
Ronin15Compare SDL3 HammerEngine metrics against baselines to flag performance regressions.
open-nekoCreate and iteratively improve AI agent skills through structured evaluation and benchmarking.
susomejiasAutomate creation, testing, and optimization of AI agent skills.
