retargeting-funnel

Pass

Audited by Gen Agent Trust Hub on Sep 13, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external site traffic and conversion metrics to design marketing sequences, which introduces an inherent surface for indirect instruction injection.\n
  • Ingestion points: The skill ingests user-provided funnel reports (files or text) containing traffic volumes, conversion rates, and time-lag distributions as specified in Workflow Step 1 of SKILL.md.\n
  • Boundary markers: There are no explicit instructions within the workflow to use delimiters or ignore embedded natural language instructions within the metric data.\n
  • Capability inventory: The agent uses ingested data to generate strategic funnel plans and measurement schedules. Capability includes file reading and read-only web browsing for benchmarks (SKILL.md). No shell execution, file writing, or external network write capabilities are present.\n
  • Sanitization: No explicit sanitization or validation of the structural integrity of the input data is defined in the instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 13, 2026, 03:32 PM
Security Audit — agent-trust-hub — retargeting-funnel