ads-server-side-tracking
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted external data sources, creating a surface for indirect prompt injection.
- Ingestion points: Processes external data from debug pages, tag payloads, system logs, exports, and vendor API responses as defined in the skill body.
- Boundary markers: The skill explicitly instructs the agent to treat these sources as untrusted data and includes a directive to never retain raw identifiers or credentials in artifacts.
- Capability inventory: The auditing process implies capabilities for file reading and potentially network requests to compare browser/server events, though no specific scripts are provided in this file.
- Sanitization: The instructions mandate inspecting hashing and minimization of data before transmission, which provides a layer of protection against raw data exposure.
Audit Metadata