carbon-footprint-tracking
Pass
Audited by Gen Agent Trust Hub on Mar 16, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [PROMPT_INJECTION]: The skill contains a vulnerability surface for indirect prompt injection where the agent processes external data inputs.
- Ingestion points: Untrusted data enters the agent context through supplier names, activity descriptions, and material types within the calculator and tracker methods in
SKILL.md. - Boundary markers: The instructions do not define explicit delimiters to separate user-provided data from agent instructions.
- Capability inventory: The skill uses Python for data processing and mentions tools for network-based data collection (e.g.,
requests,beautifulsoup4). - Sanitization: No validation or sanitization of input strings is implemented prior to processing.
- [EXTERNAL_DOWNLOADS]: The skill references several well-known external sources for environmental data and accounting.
- It identifies reputable emission factor databases including the US EPA, UK DEFRA, Ecoinvent, and the GHG Protocol.
- It lists common libraries such as
requestsandbeautifulsoup4as tools for integrating with external APIs and scraping data, which are standard for climate accounting tasks.
Audit Metadata