stitch::upload-to-stitch

Upload images, HTML, and markdown files to Stitch projects via the BatchCreateScreens API.

Updated Jun 27, 2026
One-click install
npx skills add https://github.com/khulnasoft-bot/scalify --skill stitch-upload-to-stitch-khulnasoft-bot
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: stitch::upload-to-stitch
Source: https://github.com/khulnasoft-bot/scalify/tree/main/skills/plugins/stitch-design/skills/upload-to-stitch
Command: npx skills add https://github.com/khulnasoft-bot/scalify --skill stitch-upload-to-stitch-khulnasoft-bot

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires certifi, and includes scripts (resource) components.

What problem does it solve? AI models cannot upload files to Stitch through MCP tool calls because base64-encoded file content exceeds the model's output token limit (~16K tokens), causing truncated and corrupted payloads. This Skill bypasses that limit by running a Python script that reads the file locally and sends it directly over HTTP. ## Core Features & Use Cases - Direct HTTP Upload: Reads local files, base64-encodes them in-process, and posts them to the Stitch BatchCreateScreens REST endpoint with no token limits. - Multiple File Types: Supports PNG, JPG, JPEG, WebP images as screenshots, plus HTML and Markdown files as document screens with automatic MIME type detection. - API Key Discovery: Guides extraction of the Stitch API key from MCP configuration files for Antigravity, Gemini CLI, or Claude Code. - Use Case: After generating a design mockup as a PNG or extracting a static HTML page, upload it to your Stitch project as a new screen without hitting MCP output token truncation. ## Quick Start Upload the file mockup.png to my Stitch project using the upload-to-stitch skill.

Frequently Asked Questions about stitch::upload-to-stitch

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I upload images to a Stitch project?▼

Run the upload_to_stitch.py script with your project ID, file path, and API key. It base64-encodes the file in-process and posts it to the Stitch BatchCreateScreens endpoint, supporting PNG, JPG, JPEG, and WebP formats.

Why do MCP tool calls fail when uploading files to Stitch?▼

MCP tool arguments are part of the model's output, so the model must re-emit the entire base64-encoded file as generated text. Files larger than the ~16K token output limit get truncated mid-string, producing corrupted payloads the API rejects.

What file types can I upload to Stitch with this script?▼

The script supports images (.png, .jpg, .jpeg, .webp) uploaded as screenshot screens, plus HTML (.html, .htm) and Markdown (.md) files uploaded as document screens. MIME type is auto-detected from the file extension.

Where do I find my Stitch API key?▼

Check your MCP server configuration: .gemini/antigravity/mcp_config.json for Antigravity, ~/.gemini/settings.json for Gemini CLI, or ~/.claude.json for Claude Code. Extract the key from the X-Goog-Api-Key header or auth argument.

Why does the Stitch upload fail with an SSL certificate error on macOS?▼

The SSLCertVerificationError means your Python installation lacks configured root certificate authorities. Install the certifi package, which the script uses automatically, or set the SSL_CERT_FILE environment variable to certifi's CA bundle path.