stitch-vue-vant-components
Pass
Audited by Gen Agent Trust Hub on Jun 13, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses a bash script
scripts/fetch-stitch.shto download design assets. The script executescurlwith user-provided or tool-provided URLs. While the script uses best practices like variable quoting andset -e, it represents a command execution surface. - [EXTERNAL_DOWNLOADS]: The skill fetches HTML and metadata from external Google Cloud Storage URLs using the
fetch-stitch.shscript. These URLs are retrieved dynamically from thestitch-mcp-get-screentool. - [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection (Category 8). It ingests untrusted HTML data downloaded from external sources and instructs the agent to parse and convert this content into Vue components. Maliciously crafted design data could attempt to influence the agent's output.
- Ingestion points:
temp/source.htmldownloaded viascripts/fetch-stitch.sh. - Boundary markers: Absent. The skill does not define specific delimiters or instructions to ignore embedded commands in the source HTML.
- Capability inventory:
Write(creating component files),Bash(running the fetch script andnpm install). - Sanitization: Absent. There are no instructions for the agent to sanitize or validate the content of the downloaded HTML before parsing.
Audit Metadata