Creates new Claude Code skill directories with SKILL.md files following ai-lab conventions.
npx skills add https://github.com/PatriotAi/ai-lab --skill skill-new-patriotai
Please help me install this Agent Skill. Skill: skill-new Source: https://github.com/PatriotAi/ai-lab/tree/main/.claude/skills/skill-new Command: npx skills add https://github.com/PatriotAi/ai-lab --skill skill-new-patriotai
.claude/skills/<name>/SKILL.md from the templates/skill.md template. - Convention Enforcement: Ensures minimal frontmatter (name + trigger-focused description), Ukrainian-canonical documentation with English key terms, and one-responsibility-per-skill scoping. - Use Case: A user asks to create a skill for Ukrainian-English translation; the Skill produces .claude/skills/translate-uaen/SKILL.md with a translation-trigger description, optionally catalogs it in skills/README.md, and proposes a small input-output test. ## Quick Start Ask the assistant to create a new skill for a repeated task, providing its name, purpose, and when it should trigger.High-intent search queries and answers about installing and using this skill.
Define a kebab-case name, a one-sentence purpose, and a clear trigger condition, then create .claude/skills/<name>/SKILL.md following the templates/skill.md template. The frontmatter needs only name and description, where description states when the skill activates.
Under the ai-lab lightweight standard, SKILL.md requires only name and description in its YAML frontmatter. The description must be a concrete trigger so the skill is auto-invoked correctly; vague descriptions cause missed or wrong activations.
New skills go in .claude/skills/<name>/SKILL.md so Claude Code picks them up automatically. If the skill is worth cataloging, it should also be referenced in skills/README.md.
Skills belonging to the melania ecosystem follow its full requirements (extended frontmatter, evals, guard) via skill-creation-guide instead. The lightweight ai-lab standard is meant for simple laboratory skills only.
Auto-invocation depends entirely on the description field in the frontmatter. If the trigger wording is vague or does not match how users phrase requests, the skill will not activate or will fire in the wrong context.
practical-stackExtract reusable patterns from completed session transcripts into formal feature requests.
zkysar1Evaluate agent learning state against a verification checklist using evidence from logs and journals.
artusjsAutomate drafting, evaluating, and optimizing Claude Code skills with SKILL.md validation.
Ahoo-WangCreate and evaluate Claude skills with SKILL.md and eval workflows.