skills/datadrivenconstruction/ddc_skills_for_ai_agents_in_construction/n8n-photo-report/Gen Agent Trust Hub
n8n-photo-report
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill includes a Python helper script (
SKILL.md) designed to read images from the local filesystem and upload them to a network endpoint via HTTP POST. This facilitates the intended photo reporting automation.\n- [INDIRECT_PROMPT_INJECTION]: The n8n workflow definition uses AI-powered image analysis via the Anthropic Claude Vision API. This creates a potential surface for indirect prompt injection, where content within processed images could theoretically influence the AI's behavior or output.\n - Ingestion points: Image data processed by the 'Analyze with Claude Vision' node in 'SKILL.md'.\n
- Boundary markers: The prompt instructions for the vision model do not include specific delimiters or instructions to ignore embedded content.\n
- Capability inventory: The skill possesses filesystem read permissions and network communication capabilities through the Python helper script and n8n workflow configuration.\n
- Sanitization: The skill does not explicitly describe sanitization or filtering logic for the visual content of processed images.
Audit Metadata