customer-discovery
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill includes a Python script (
scripts/main.py) that functions as a command-line interface for managing discovery data. It performs standard file operations, reading from and writing to a localcustomer_discovery.jsonfile to store hypotheses and interview notes. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest untrusted data from the user, creating a surface for indirect prompt injection.
- Ingestion points: The agent is instructed to collect business ideas, hypotheses, and detailed customer interview summaries from the user in
SKILL.md. - Boundary markers: There are no explicit instructions or delimiters defined to prevent the agent from potentially following instructions embedded within the user-supplied interview data.
- Capability inventory: The skill has the capability to write the collected data to the local filesystem using the provided Python script (
scripts/main.py). - Sanitization: The skill does not implement validation or sanitization of the input text before processing or persistent storage.
Audit Metadata