skills/aibangjuxin/knowledge/ob/Gen Agent Trust Hub

ob

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFEPROMPT_INJECTIONNO_CODE
Full Analysis
  • [PROMPT_INJECTION] (LOW): Detected a potential surface for Indirect Prompt Injection (Category 8).
  • Ingestion points: The skill parses external YAML/JSON configurations and GKE/Cloud Logging entries as part of its onboarding workflow.
  • Boundary markers: Absent; there are no instructions to use delimiters or ignore embedded commands in the processed data.
  • Capability inventory: None; the skill is purely instructional and contains no scripts, binaries, or tool-calling definitions that could be exploited.
  • Sanitization: Absent; no mention of sanitizing or validating the input data against malicious content.
  • [NO_CODE] (SAFE): The skill consists entirely of Markdown text and YAML metadata. No scripts, binaries, or automated execution patterns were found.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:33 PM