sales-performcb

Pass

Audited by Gen Agent Trust Hub on Apr 20, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill functions as a comprehensive reference for the Perform[cb] platform, covering campaign management, fraud detection (PerformSHIELD), and API integration. No executable scripts or automated network operations were found within the skill files.
  • [INDIRECT_PROMPT_INJECTION]: The skill uses a 'self-improving' mechanism by reading from and writing to references/learnings.md. This represents a potential surface for indirect prompt injection where untrusted data from user sessions could be persisted. However, the skill lacks any high-risk capabilities (such as subprocess execution, file system modification outside its own folder, or outbound network tools) that could be leveraged by such an injection.
  • Ingestion points: User input during interactions and the references/learnings.md file itself.
  • Boundary markers: Absent; there are no specific delimiters to separate user-provided content from instructions.
  • Capability inventory: No tool usage, subprocess calls, or network operations are defined in the skill scripts.
  • Sanitization: None; user findings are appended directly to the reference file.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 20, 2026, 06:35 PM