outbound-optimizer

Pass

Audited by Gen Agent Trust Hub on Mar 24, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes untrusted external data.
  • Ingestion points: Step 1 explicitly asks for a "Sample: Best-performing email or script" from the user or the environment.
  • Boundary markers: The skill lacks explicit instructions to isolate this untrusted content (e.g., wrapping in XML tags or using specific delimiters with instructions to ignore embedded commands).
  • Capability inventory: The skill has access to sensitive tools including Read, Write, Edit, WebSearch, and WebFetch. A malicious payload in an email sample could attempt to manipulate the agent into misusing these tools.
  • Sanitization: No sanitization, validation, or filtering of the provided email samples is defined.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 24, 2026, 09:57 PM