Help Center · Skills
Skills
Author SKILL.md bundles, lint and critique against six dimensions, and run multi-provider Behavioral Eval. Import from Anthropic SKILL.md, OpenAI Custom GPTs, or Gemini Gems.
- 01Skills overviewWhat a Skill is in Prompt Assay, how it differs from a prompt, and the workbench surfaces that author + critique + evaluate one.
- 02Behavioral Eval guideAuthor trigger + non-trigger probes, run them across 2-5 models in parallel, and read the activation + adherence + cross-provider consistency scores.
- 03README badgeEmbed a Shields.io badge in your skill's README that links to its public Skill Report.
- 04Authoring a Skill from scratchFrontmatter contract, body conventions, scripts/ vs references/ vs body, and the common pitfalls that trip the linter on the first save.
- 05Sharing Skill Reports publiclyWhat gets published, the body-publish opt-in (authorship-gated), privacy defaults, view counting, revocation, and the carve-out from PA's no-server-side-AI-cache rule.
- 06Tiers, cost, and BYOK for SkillsPer-tier caps for Behavioral Eval, what each Skill instrument bills your provider, the $5 cost-preflight gate, and how to read the per-skill cost dialog.