research-audience-score-segment-qa
Pass
Audited by Gen Agent Trust Hub on Nov 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external, untrusted data that could contain malicious instructions designed to manipulate the agent's logic or scoring outcomes.
- Ingestion points: The agent reads and processes data from the
account table,people rows, andsignalsinSKILL.md. It also ingests text from a quote bank (voc-bank.md) and analyzes external content like "job posts", "case studies", and "integration pages" to perform competitive analysis. - Boundary markers: The instructions do not define any delimiters or clear boundary markers to help the agent distinguish between its operational instructions and the text content it is analyzing.
- Capability inventory: The skill possesses the capability to write sensitive lead data to the local file system (
list.csv,segments.csv), access marketing databases (marketing-data-access), and store information in long-term memory (save-to-brain,save-learning). - Sanitization: No mechanisms for sanitizing or filtering the input text for prompt injection patterns are described in the instructions.
Audit Metadata