Releases: Do-fei/website-studio-skill
Releases · Do-fei/website-studio-skill
Release list
Website Studio v1.0.0
Website Studio v1.0.0
Website Studio turns a website request into a complete, visually intentional, responsive, and honestly verified deliverable. It is a portable Agent Skill rather than a fixed template or framework scaffold.
Included in this release
- Seven-gate workflow: brief, design direction, product shape, build, capabilities, verification, and publishing handoff.
- Eight website recipes: landing page, portfolio, editorial story, knowledge hub, directory, dashboard, portal, and tracker.
- Four capability recipes: data, authentication, uploads, and external connectors.
- Four execution adapters: OpenAI Sites, existing projects, static sites, and generic frameworks.
- Six deterministic validators and ten automated tests.
- Three original responsive HTML examples.
- Detailed installation guidance for Codex, Claude Code, Gemini CLI, GitHub Copilot, Cursor, Windsurf, OpenCode, Cline, Roo Code, and generic file-aware agents.
Install
The Release asset contains a single website-studio/ folder. Extract that folder into your Agent's Skill discovery directory.
The broad cross-Agent project location is:
.agents/skills/website-studio/
Claude Code uses .claude/skills/website-studio/; Cline most reliably uses .cline/skills/website-studio/. See the repository's cross-Agent guide for all personal and project paths.
Verification
- Package structure: passed, 66 required files.
SKILL.md: 83 lines with progressive disclosure.- Automated tests: 10/10 passed.
- Script syntax: passed.
- SVG/XML validation: passed.
- Release ZIP: extracted into a fresh directory and
npm testpassed from the extracted package. - No bundled binary media or third-party brand assets.
Release asset integrity
SHA-256 710f847c930b6b0644f91603efb74b5c8fcfda5a75c3ccf7f74a2eaf8bedfd37
File website-studio-v1.0.0.zip
Review SKILL.md and scripts/ before granting an Agent terminal or deployment permissions. Hosting a Skill on GitHub does not make it automatically trusted.