181-java-observability-logging
Pass
Audited by Gen Agent Trust Hub on Sep 26, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides detailed guidance on secure logging practices, including explicit instructions to avoid logging sensitive data (passwords, tokens, secrets) and techniques for masking data like credit card numbers.
- [SAFE]: Log injection prevention is addressed by recommending the sanitization of untrusted text before logging to prevent potential log forging attacks.
- [SAFE]: The workflow incorporates mandatory safety checks using standard build tools (mvn compile and mvn clean verify) to ensure project stability and code integrity before and after applying logging changes.
- [INDIRECT_PROMPT_INJECTION]: The skill represents a surface for indirect prompt injection as it processes untrusted source code and possesses file-write and command-execution capabilities. 1. Ingestion points: The agent reads Java project source code and configuration files (e.g., logback.xml) to assess and refactor logging. 2. Boundary markers: There are no explicit boundary markers or instructions to ignore embedded commands in the processed data. 3. Capability inventory: The skill utilizes file-write access to update logging code and shell execution for Maven build commands. 4. Sanitization: The skill does not explicitly define sanitization of input data before prompt interpolation, although it instructs the agent to implement sanitization in the resulting code.
Audit Metadata