skills/reason-machines/marketing-skills/genpark-marketing-cloud-sql-audience-generator/Gen Agent Trust Hub
genpark-marketing-cloud-sql-audience-generator
Fail
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: HIGHEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill documentation directs users to clone a repository from an unverified GitHub account (alphaparkinc) which is not recognized as a trusted vendor or well-known service in the context of the author.
- [REMOTE_CODE_EXECUTION]: The provided installation instructions include running
pip install -r requirements.txton the unverified external repository, which could lead to the execution of malicious code during the package installation process. - [INDIRECT_PROMPT_INJECTION]: The skill facilitates the conversion of untrusted natural language intents into SQL queries for execution in Salesforce Marketing Cloud. This architectural pattern is vulnerable to prompt injection where a crafted input could result in harmful SQL generation.
- Ingestion points: The
intentparameter in thegenerate_sqlfunction withinSKILL.md. - Boundary markers: No delimiters or "ignore" instructions are used to isolate the user-provided intent from the system prompt.
- Capability inventory: The skill possesses the capability to execute generated SQL against production data extensions via the
MarketingCloudClient. - Sanitization: There is no evidence of input sanitization or query parameterization to prevent SQL injection or logic manipulation.
Recommendations
- AI detected serious security threats
Audit Metadata