churn-prevention
Warn
Audited by Gen Agent Trust Hub on Apr 5, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute a local script that is not provided in the bundle, making its behavior unverifiable.\n
- Evidence: The file
SKILL.mdcontains a shell commandpython3 scripts/churn_impact_calculator.py.\n - Evidence: The "Output Artifacts" section explicitly directs the agent to "Run churn_impact_calculator.py with your inputs" when modeling churn impact.\n- [PROMPT_INJECTION]: The skill contains an attack surface for indirect prompt injection due to unvalidated data ingestion.\n
- Ingestion points: The agent is instructed to read
marketing-context.md(if present) before beginning its task, as noted inSKILL.md.\n - Boundary markers: There are no specified delimiters or instructions to treat the content of
marketing-context.mdas untrusted data.\n - Capability inventory: The agent is empowered to execute shell commands and perform business logic calculations.\n
- Sanitization: No sanitization or validation logic is defined for the external context being ingested.
Audit Metadata