debugging-techniques

Pass

Audited by Gen Agent Trust Hub on Apr 29, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious patterns, obfuscation, or unauthorized data access were detected. The skill focuses on legitimate administrative and development tasks within the ServiceNow platform and uses authorized tools for its stated purpose.- [PROMPT_INJECTION]: The skill identifies a potential surface for indirect prompt injection because it instructs the agent to read external data from ServiceNow system logs and tables. However, this is consistent with its purpose as a debugging tool.
  • Ingestion points: The SN-Query-Table tool is used to ingest data from tables such as syslog and sys_script_execution_history (SKILL.md).
  • Boundary markers: No explicit delimiters or instructions to ignore embedded content are provided in the examples.
  • Capability inventory: The skill utilizes SN-Execute-Background-Script and Bash, which provide significant interaction capabilities with the instance and environment (SKILL.md).
  • Sanitization: No sanitization or filtering logic is described for the content retrieved from the logs.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 29, 2026, 04:35 PM