What problem does it solve? Authoring agent skills by hand leads to bloated SKILL.md files, missing discovery metadata, namespace collisions with other skill suites, and inconsistent packaging across harnesses like Claude Code, Codex, and Gemini. ## Core Features & Use Cases - Guided Skill Authoring: Interactive intake workflow that gathers name, purpose, triggers, and boundaries, then generates a concise SKILL.md orchestrator with one-level modular references, examples, and eval checklists. - Collision and Clobber Guards: Runs namespace collision checks against shared global skill roots and a clobber guard before any write, preventing overwrites of unrelated same-named skills. - Cross-Harness Packaging: Packages a canonical skill into an Anthropic-schema-compliant .skill zip for claude.ai upload, bundling cross-repo references and trimming frontmatter automatically. - Use Case: Ask to create a new orama-system sub-skill for rotating API keys; skillify interviews you for missing facts, previews the frontmatter, writes SKILL.md plus references and eval files, validates line counts and code fences, and registers it only after your confirmation. ## Quick Start Ask the agent to create a new orama-system skill for your task, specifying the target harness and trigger phrases, then confirm the proposed outline before any files are written.
