road-trip-planner
Pass
Audited by Gen Agent Trust Hub on Oct 11, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external web searches to generate itineraries, creating a surface for indirect prompt injection. \n
- Ingestion points: Web search results for destination facts, operating hours, and weather (SKILL.md, Steps 3-5). \n
- Boundary markers: Absent; the skill does not define delimiters for external data. \n
- Capability inventory: File system write, web search (SKILL.md, Steps 6-9). \n
- Sanitization: Absent; no specific sanitization logic is provided for searched content.
- [DYNAMIC_EXECUTION]: The skill uses dynamic execution to generate visual summary cards. \n
- Evidence: Step 8 describes generating a Python script using the Pillow library or using a headless browser to capture a screenshot of a dynamically generated HTML file.
- [EXTERNAL_DOWNLOADS]: The skill references the use of external packages from well-known sources for utility tasks. \n
- Evidence: The workflow suggests using 'uvx playwright' to generate PNG cards, which involves executing a remote package from an established registry. This is documented as part of standard functionality.
Audit Metadata