agentleFS
Sign inStart free
PracticalSwanPracticalSwan/agent-skills/stitch-loop/SKILL.md

stitch-loop

Skill22 starsChanged 24 days agoMITView on GitHub

Run an iterative Stitch website-building loop using `.stitch/next-prompt.md`, SITE.md, DESIGN.md, generated pages, and verification checkpoints.

12 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use find-skills.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • VercelSkill

    find-skills

    Helps users discover and install agent skills when they ask questions like "how do I do X", "find a skill for X", "is there a skill that can...", or express interest in extending capabilities. This skill should be used when the user is looking for functionality that might exist as an installable skill.

    Also in 3 projects that use this

The file

We split this skill into the rules it gives an agent

11 rules. 0 are common, 6 are distinctive, and 5 are very specific to this project.

  • 0

    Common

    Most of the 9 edited versions we read keep these.

  • 6

    Distinctive

    This author's own choices. Worth a look.

  • 5

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use when a multi-page site should be built iteratively with Stitch and baton files.

    line 15
  2. Very specific:

    Read .stitch/next-prompt.md, .stitch/SITE.md, and .stitch/DESIGN.md before generating or integrating pages.

    line 21
  3. Very specific:

    Use stitch-generate-design for screen prompts when tools or the Stitch web UI are available.

    line 22
  4. Distinctive:

    When screen-generation MCP tools are absent, prepare the next prompt and record that the web UI/API step is required.

    line 23
  5. Distinctive:

    Run local browser verification when a page is integrated into a runnable site.

    line 25
  6. Very specific:

    Update SITE.md, metadata, and next-prompt.md before closing the loop.

    line 26
  7. Very specific:

    Keep generated .stitch/ files out of commits unless the user explicitly wants them as durable examples.

    line 32
  8. Very specific:

    Use local .stitch/ artifacts, exported HTML or screenshots, bundled scripts, and the Stitch web UI when the host does not expose the needed …

    line 58
  9. Distinctive:

    Do not claim screen lookup, generation, editing, or variant MCP calls unless those tools are present in the active host tool list.

    line 59
  10. Distinctive:

    Do not claim an MCP operation was used when the active host does not expose it.

    line 60
  11. Distinctive:

    Before claiming this skill was applied successfully:

    line 74

“Common” is measured: kept by at least 6 in 10 of the 9 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Stitch Build Loop

This skill is a catalog-normalized import from https://github.com/google-labs-code/stitch-skills at commit 7b53207b94e62911777d53d4238b5f8c88c2b519, source path plugins/stitch-utilities/skills/stitch-loop. The upstream control file was corrected for this workspace: the verified Stitch MCP surface here is design-system oriented, so screen lookup, screen generation, and screen editing tools must be used only when the current host explicitly exposes them.

When to Use This Skill

  • Use when a multi-page site should be built iteratively with Stitch and baton files.
  • The task involves Google Stitch project IDs, .stitch/ artifacts, DESIGN.md files, Stitch exports, or Stitch-specific validation.
  • The broader stitch-design router points here as the narrowest workflow.

Workflow

  1. Read .stitch/next-prompt.md, .stitch/SITE.md, and .stitch/DESIGN.md before generating or integrating pages.
  2. Use stitch-generate-design for screen prompts when tools or the Stitch web UI are available.
  3. When screen-generation MCP tools are absent, prepare the next prompt and record that the web UI/API step is required.
  4. Move generated or exported HTML into the site only after checking the sitemap.
  5. Run local browser verification when a page is integrated into a runnable site.
  6. Update SITE.md, metadata, and next-prompt.md before closing the loop.

Local Assets

  • examples/, resources/, references/, or reference/ are upstream support material when present. Treat SKILL.md as the source of truth if a support file mentions an unavailable MCP tool.
  • scripts/ are optional helpers. On Windows, prefer PowerShell or Node equivalents unless Git Bash or WSL is actually available.
  • Keep generated .stitch/ files out of commits unless the user explicitly wants them as durable examples.

Corrected Stitch MCP Surface

Verified in this workspace on 2026-06-15: create_project, upload_design_md, create_design_system_from_design_md, list_design_systems, and apply_design_system. This 2026-07-29 source refresh did not re-verify a broader live MCP surface. Do not claim list_projects, list_screens, get_project, get_screen, generate_screen_from_text, edit_screens, or generate_variants were used unless the current host exposes those exact tools in the active tool list.

<!-- MCP:START --><!-- PORTABILITY:START -->

Cross-Client Portability

This skill is written to stay usable across GitHub Copilot, Claude Code, and Codex.

  • GitHub Copilot: keep the folder in a Copilot-visible skill path or wrap the workflow in project instructions when folder discovery is unavailable.
  • Claude Code: keep the folder in a local skills directory or a compatible plugin source.
  • Codex: install or sync the folder into $CODEX_HOME/skills/stitch-loop and restart Codex after major changes.
<!-- PORTABILITY:END -->

MCP Availability And Fallback

Preferred MCP Server: Stitch MCP

  • Fallback prompt: "Use the Stitch Build Loop skill without MCP. Follow the documented local or manual fallback, show the selected tool surface, and report the verification evidence."
  • Use local .stitch/ artifacts, exported HTML or screenshots, bundled scripts, and the Stitch web UI when the host does not expose the needed Stitch MCP operation.
  • Do not claim screen lookup, generation, editing, or variant MCP calls unless those tools are present in the active host tool list.
  • Do not claim an MCP operation was used when the active host does not expose it.
<!-- MCP:END -->

Anti-Patterns

  • Claiming a Stitch screen-generation, screen-editing, or screen-retrieval MCP call succeeded when the active host does not expose that tool.
  • Uploading files, screenshots, HTML, markdown, or design assets to Stitch without user-approved destination and artifact details.
  • Reading, printing, storing, or committing Stitch API keys, MCP config secrets, cookies, or credential-bearing files.
  • Treating generated design or code as final without local render, syntax, or artifact verification.
  • Collapsing this workflow into a broader frontend/design skill when Stitch-specific files, project IDs, or design-system assets matter.

Verification Protocol

Before claiming this skill was applied successfully:

  1. Pass/fail: Baton frontmatter has a page slug and concrete task body.
  2. Pass/fail: SITE.md reflects sitemap, completed pages, and next backlog item.
  3. Pass/fail: Generated or integrated pages were checked locally or unavailable generation is recorded.
  4. Pass/fail: The loop leaves a next baton for the following session.
  5. Pressure-test scenario: Repeat the workflow with Stitch MCP screen tools unavailable and confirm the fallback path remains honest and actionable.
  6. Success metric: The user can identify the exact artifact, project/design-system target, and verification evidence without relying on unstated MCP behavior.

Related Skills

  • stitch-generate-design: Use when the task also needs this adjacent Stitch workflow.
  • stitch-design-md: Use when the task also needs this adjacent Stitch workflow.
  • stitch-extract-static-html: Use when the task also needs this adjacent Stitch workflow.
  • web-testing: Use when the task also needs this adjacent Stitch workflow.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free