logging-patterns

Pass

Audited by Gen Agent Trust Hub on Mar 5, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent on using shell commands such as cat, tail, jq, and mvnw to run the application and process log data.
  • [INDIRECT_PROMPT_INJECTION]: The skill facilitates the analysis of application logs, which constitutes an ingestion point for untrusted data.
  • Ingestion points: Application log files (e.g., app.log) which record data from external requests, such as userId, message contents, or X-Request-ID headers.
  • Boundary markers: The skill recommends using JSON (structured logging), which helps separate data fields from instructions, but does not implement specific delimiters or 'ignore' instructions for the agent when reading log content.
  • Capability inventory: The skill utilizes shell-based log processing tools (jq, cat, tail) and application execution via Maven (mvnw).
  • Sanitization: The skill explicitly advises against logging sensitive information like passwords, PII, and tokens, which reduces data exposure risk but does not filter for potential prompt injection strings within logged messages.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 5, 2026, 12:27 PM