analytics
Pass
Audited by Gen Agent Trust Hub on Jun 14, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill attempts to ingest data from untrusted local context files which presents a risk of indirect instruction injection.
- Ingestion points: The skill instructs the agent to read .agents/product-marketing.md, .claude/product-marketing.md, or product-marketing-context.md (documented in SKILL.md).
- Boundary markers: No explicit delimiters or instructions to ignore embedded commands are specified for these context files.
- Capability inventory: The skill is used to generate implementation code (JavaScript snippets for GA4 and GTM) and provide strategic advice based on these inputs.
- Sanitization: No validation or sanitization of the input file content is mentioned in the skill instructions.
- [EXTERNAL_DOWNLOADS]: The skill provides documentation and implementation examples that reference scripts from well-known technology services.
- Evidence: The Google Tag Manager reference (references/gtm-implementation.md) includes a standard Facebook Pixel snippet that loads fbevents.js from connect.facebook.net.
- Context: This is a standard practice for analytics implementation and involves a well-known service provider.
Audit Metadata