Draft SKILL.md files and benchmark trigger accuracy with eval runs.
npx skills add https://github.com/XCmiaow/cc-deploy --skill skill-creator-xcmiaow
Please help me install this Agent Skill. Skill: skill-creator Source: https://github.com/XCmiaow/cc-deploy/tree/main/skills/skill-creator Command: npx skills add https://github.com/XCmiaow/cc-deploy --skill skill-creator-xcmiaow
This Skill helps you design new Claude Code skills or iteratively improve existing ones by running structured trigger evaluations and using the results to refine the skill description for better accuracy.
Use the skill to create a new skill by telling it what you want your skill to do and how you want users to recognize when it should trigger.
High-intent search queries and answers about installing and using this skill.
Improve Claude Code skill triggering accuracy by running an eval-driven loop that generates realistic test queries, executes with-skill and without-skill baselines, and uses automated benchmarking to iteratively refine the SKILL.md description.
An eval-driven loop for prompt engineering is a structured iterative process where you create test cases, run benchmark evaluations against baselines, and select improved descriptions while avoiding overfitting to narrow examples.
Create a new Claude Code skill from scratch by defining your desired workflow, drafting the SKILL.md instructions, generating trigger evals, and refining the description based on quantitative benchmark reviews.
You can refactor an existing Claude Code skill description by running trigger evaluations, inspecting performance patterns in a review UI, and testing alternative descriptions to select the best-performing one without overfitting.
Your Claude Code skill may trigger incorrectly due to an overly broad description, which you can fix by generating a test set of real user queries and running automated benchmarks to isolate and optimize the triggering behavior.
Pryma-TechExtracts workflows from sessions and drafts reusable SKILL.md documents with placement guidance.
niduran-orionCreate, evaluate, and refine Claude skills with SKILL.md and trigger evaluations.
slabgorbRun predefined benchmarks comparing persona effectiveness across code-review, test-writing, and architecture tasks.
pferretti99Create or refine Claude skill directories with SKILL.md and evaluation loops.
