skills/datadrivenconstruction/ddc_skills_for_ai_agents_in_construction/data-anomaly-detector/Gen Agent Trust Hub
data-anomaly-detector
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied data in formats like Excel, CSV, and JSON, which creates a vector for instructions embedded within the data to influence agent behavior.
- Ingestion points: Data is loaded via pandas from project files as seen in SKILL.md and instructions.md.
- Boundary markers: The instructions lack explicit delimitation or warnings to the agent to disregard instructions found within the data files.
- Capability inventory: The skill utilizes filesystem access for reading project data.
- Sanitization: No sanitization or validation logic is implemented to strip potential prompt injection payloads from the construction data.
Audit Metadata