Create and refine Agent Skills with SKILL.md and eval generation.
npx skills add https://github.com/Yukine5377/common --skill skill-creator-yukine5377
Please help me install this Agent Skill. Skill: skill-creator Source: https://github.com/Yukine5377/common/tree/main/.cursor/skills/skill-creator Command: npx skills add https://github.com/Yukine5377/common --skill skill-creator-yukine5377
This Skill provides a structured, end-to-end process to design, implement, evaluate, and iteratively improve Agent Skills. It guides you from capturing user intent and drafting SKILL.md to crafting test prompts, running evaluations, visualizing results, and refining the description until activation reliability is high.
Provide your initial skill idea; this tool scaffolds SKILL.md, generates a basic eval set, and starts the iterative loop toward a refined skill.
High-intent search queries and answers about installing and using this skill.
Agent Skill creation involves capturing user intent, drafting a SKILL.md file, generating test prompts, running evals, and iteratively refining descriptions to maximize activation reliability and overall usefulness.
Evaluating Agent Skills involves running test prompt sets, collecting performance metrics, and visualizing benchmarking results to measure activation reliability and behavioral accuracy.
Improving activation reliability requires iteratively refining the skill description and test coverage based on evaluation results, repeating the benchmarking loop until the Agent consistently triggers the correct skill.
Yes, the skill-creator workflow relies on the anthropic and pyyaml dependencies to scaffold SKILL.md files, package metadata, and execute scripts for automated testing and evaluation.
The best way to structure test prompts is to fabricate an initial eval set that probes specific skill behaviors, then expand test coverage based on benchmarking results and iterative refinement.
AzrinPutraGuide exploratory conversations to elicit project requirements and generate an intent brief.
hashintelGenerate and validate SKILL.md frontmatter and guidance for agent skills.
HemantSudarshanAutomate evaluation and benchmarking of AI agents and LLM-powered systems.
yasuwrldhyperCreate and refine AI skills with SKILL.md metadata and benchmark testing.
