business-decomposition-order-quality

Pass

Audited by Gen Agent Trust Hub on Sep 19, 2026

Risk Level: SAFE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes external data from financial disclosure platforms and procurement websites.
  • Ingestion points: External data from sources like CNINFO and operator procurement portals (specified in SKILL.md and references/idc-order-ledger-guide.md).
  • Boundary markers: None present; the skill lacks explicit instructions to ignore embedded commands in the source data.
  • Capability inventory: No dangerous capabilities such as file system writes, network exfiltration, or code execution were detected.
  • Sanitization: No input sanitization or validation logic is defined.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 19, 2026, 02:21 AM
Security Audit — agent-trust-hub — business-decomposition-order-quality