amazon-ads-marketing-cloud
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from external sources, specifically Amazon Marketing Cloud (AMC) APIs and CSV results. This creates an indirect prompt injection surface where external data could contain malicious instructions. The skill mitigates this risk by explicitly instructing the agent to summarize results and never pull full CSV bodies into the model context.
- [DYNAMIC_EXECUTION]: The skill provides Python templates for processing CSV data and instructs the agent to use an available execution facility to run these snippets. The snippets use standard libraries (statistics, csv, io) to perform analytical tasks like calculating weighted averages or percentiles, which is a common and legitimate pattern for data analysis skills.
- [EXTERNAL_DOWNLOADS]: The skill facilitates downloading analysis results from Amazon S3 using pre-signed URLs provided by the AMC API. These downloads target a well-known service (AWS) and are part of the core functionality of the skill, posing no risk in this context.
Audit Metadata