arcgis-starter-app
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to scaffold project files using user-provided strings (such as project names, map IDs, and API keys) which presents a potential surface for indirect injection if these strings contain malicious instructions.
- Ingestion points: Project names in
package.jsonanditem-idattributes inindex.htmltemplates. - Boundary markers: None present in the provided templates to delimit user input from shell commands or code logic.
- Capability inventory: File write operations and shell command execution (
npm install,vite) required to scaffold the project. - Sanitization: The skill does not explicitly provide logic to sanitize user-provided strings before interpolation into configuration files.
Audit Metadata