sec-cf3-unsustainable-cffo

Pass

Audited by Gen Agent Trust Hub on Feb 18, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [Indirect Prompt Injection] (LOW): The skill identifies a potential attack surface through the ingestion of untrusted financial data (DPO, DIO, AR) for analysis.
  • Ingestion points: Runtime financial metrics processed via user prompts.
  • Boundary markers: Absent; the instructions do not explicitly define delimiters for untrusted data.
  • Capability inventory: None; the skill has no file-system access, network capabilities, or command execution functionality.
  • Sanitization: Absent; no explicit validation of input data is defined.
  • [No Code] (SAFE): The skill consists entirely of natural language instructions and YAML metadata without any scripts, binaries, or package dependencies, precluding most technical attack vectors.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 18, 2026, 09:15 AM