list-segmentation
Pass
Audited by Gen Agent Trust Hub on Sep 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted account enrichment data and personalization hooks that are designed to be consumed by downstream email generation and prompt building skills.
- Ingestion points: The skill reads account data from JSON files located in
campaigns/{slug}/accounts/, which contains externally sourced enrichment data. - Boundary markers: The instructions do not define delimiters or explicit 'ignore embedded instructions' markers to prevent the agent or downstream models from interpreting malicious instructions hidden within account fields.
- Capability inventory: The skill has file read and write capabilities, specifically reading multiple account files and writing a unified
segmentation.jsonfile. - Sanitization: There is no mention of sanitizing or escaping the
hook_qualitycontent orenrichment columnsbefore they are formatted into the output JSON and summary report.
Audit Metadata