Automate skill development from intent capture to SKILL.md drafting and evaluations.
npx skills add https://github.com/Typeless-Git/skills --skill skill-creator-typeless-git
Please help me install this Agent Skill. Skill: skill-creator Source: https://github.com/Typeless-Git/skills/tree/main/skill-creator Command: npx skills add https://github.com/Typeless-Git/skills --skill skill-creator-typeless-git
Helps teams standardize and accelerate the creation, refinement, and evaluation of Claude skills by providing a repeatable workflow from intent capture to SKILL.md frontmatter and an integrated evaluation loop.
Start by outlining your skill goal, intended prompts, and evaluation criteria, then generate the initial SKILL.md to begin the draft-and-test loop.
High-intent search queries and answers about installing and using this skill.
Skill development automation captures your intent, drafts the SKILL.md file with required frontmatter fields like name and description, and triggers evaluations to standardize the creation workflow. It conducts interviews to refine scope and success criteria.
Benchmarking skill performance involves running end-to-end evaluations and generating iteration-based workloads to test improvements. You can visualize the benchmark results across iterations using the integrated eval viewer to track progress.
An evaluation loop improves skill iteration by generating test workloads that measure deterministic execution against defined success criteria. It automates the draft-and-test cycle, allowing you to benchmark and optimize triggering descriptions continuously.
Yes, deterministic skill execution requires frontmatter fields including name and description. You can also bundle optional resources like scripts, references, and assets to support deterministic tasks and provide richer context for the workflow.
Yes, you can bundle optional resources such as scripts, references, and assets alongside your SKILL.md. These bundles support deterministic task execution and supply richer context during the end-to-end skill development and evaluation process.
Automated skill refinement workflows are not ideal when you lack a defined workflow or clear evaluation criteria. Without established intent, success metrics, or the need for benchmarking, manual drafting may be more suitable than the automated draft-and-test loop.
s977043Design a Claude Code skill from a use case and produce a repo-ready package.
cenjieAutomate Claude skill creation, SKILL.md drafting, and iterative evaluations.
103492-MaxPaardekamDesign, document, evaluate, and iteratively improve Claude skills from draft to deployment.
t0labCreate SKILL.md files with YAML frontmatter and evaluate Claude Skills iteratively.
