stitch-uview-plus-components
Pass
Audited by Gen Agent Trust Hub on Jun 13, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes HTML content fetched from an external source (Stitch design URLs) to generate code. This creates an attack surface for indirect prompt injection, where instructions embedded in the design metadata or HTML could potentially influence the agent's behavior.
- Ingestion points:
temp/source.html(populated viascripts/fetch-stitch.shand used in code generation). - Boundary markers: No explicit delimiters or boundary markers are used to isolate the ingested HTML content from the agent's instructions.
- Capability inventory:
Bash,Write,web_fetch(allowing for file modification and network access). - Sanitization: No sanitization or filtering of the ingested HTML content was detected.
- [EXTERNAL_DOWNLOADS]: The skill includes a helper script that downloads HTML and screenshots from the Stitch design platform.
- Evidence:
curl -L -f -sS --connect-timeout 10 --compressed "$URL" -o "$OUTPUT"inscripts/fetch-stitch.sh. - Context: Fetches design assets from Google's Stitch service (
stitch.withgoogle.com). - [COMMAND_EXECUTION]: The skill executes a local shell script to manage asset downloads.
- Evidence:
bash scripts/fetch-stitch.sh "<htmlCode.downloadUrl>" "temp/source.html"inSKILL.md.
Audit Metadata