ecommerce-text-trademark-detector
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes 'productTitle' and 'productText' which are user-controllable inputs. If these inputs contain malicious instructions, the agent may execute them when processing the search results returned by the API.
- Ingestion points: The 'productTitle' and 'productText' parameters processed by 'SKILL.md' and 'scripts/text_trademark_detection.py'.
- Boundary markers: Absent. The instructions do not direct the agent to ignore instructions embedded in the product data or results.
- Capability inventory: The skill script performs network POST requests and writes JSON files to the local project directory.
- Sanitization: The script uses 'json.loads' to handle the input format but does not filter or sanitize the textual content for potential prompt injection attacks.
Audit Metadata