Skip to content
SkillModel-led2,589 tokens in use

stitch-loop

Builds a website page by page with an autonomous baton loop that generates each page in Stitch, wires it into the site and writes the next task.

10,975 stars on the repository Apache License 2.0google-labs-code/stitch-skills · plugins/stitch-utilities/skills/stitch-loop
npx skills add google-labs-code/stitch-skills
View source

Install command as written in the repository's README

Skill label

measured from its files

What installing stitch-loop costs

23tokens idle
2,589tokens in use

Idle is what sits in context while the skill is installed. In use is its instructions, loaded when it runs. Counts are approximate, made offline with cl100k_base.

Model dependence Model-led

Script-ledMixedModel-led

Mostly instructions, so how well it works depends on the model that follows them.

This describes the skill's structure, how much of it is scripts. It does not say how well it works.

Touches

shelldeclared and detectedwrites filesdeclared

Declared tags come from the skill's frontmatter; detected ones from its scripts, its code blocks and the commands it names in inline code. This is detected, not audited.

How it scores

Quality

How well the skill is written, counted from its own files: its frontmatter, its steps, examples, stated output and constraints, its size and whether every file it ships is one it names.How we score it →
Good
  • Frontmatter8/15
  • Structure25/25
  • Examples12/15
  • Output15/15
  • Size15/15
  • Resources0/15

Security

Starts at 100 and loses points for each thing in the skill's files that a reader should look at first.How we score it →
Excellent

No findings.

Repository health

Pre-computed score based on development activity, maintenance, community, maturity, and trend momentum.How we score it →
Good
  • Development activity80
  • Maintenance52
  • Community68
  • Maturity32
  • Momentum40

This skill teaches an agent to run the Stitch Build Loop. Each iteration reads the task from .stitch/next-prompt.md, generates a screen with the Stitch MCP tools, saves it into the site folder, updates the sitemap and writes the next task back to the baton file.

It relies on DESIGN.md and SITE.md for style and scope, and can check the page visually through the Chrome DevTools MCP server.

What You Get

  • Generated pages saved under site/public
  • Design downloads in .stitch/designs with HTML and screenshots
  • A .stitch/metadata.json holding project and screen IDs
  • An updated SITE.md sitemap and a fresh next-prompt.md after each pass

Common Use Cases

  • Building a five-page portfolio site from a single starting prompt
  • Running page generation from CI whenever the baton file changes
  • Keeping a multi-page site consistent through one design system

What Makes It Unique

The baton file turns page generation into a relay that any orchestration layer can drive, from a human to a CI job to chained agents. Every prompt carries the design system block, so pages share one visual language.

Under The Hood

The agent calls Stitch MCP tools to create projects and generate screens, downloads HTML and screenshots, and moves files into site/public. It uses the shell to start and stop a local dev server such as npx serve when the optional Chrome DevTools MCP check is available, and it writes the baton, sitemap and metadata files.

Join founders building with open source

Opinionated takes, migration guides, cost-saving tips, and insights from the open source ecosystem.

Subscribe on SubstackJoin 750+ subscribers