clean-code-architecture
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests and processes Java source code, which constitutes untrusted data. A malicious actor could embed instructions within code comments or string literals designed to influence the agent's behavior. This risk is compounded by the skill's instruction to operate autonomously ("No human-in-the-loop") and apply changes without user confirmation, which could lead to unauthorized code modifications or the execution of unwanted follow-up tasks.
- Ingestion points: Target Java source files (controllers, services, listeners, etc.) identified through git status, diffs, or output from upstream agent skills in the pipeline.
- Capability inventory: File modification (refactoring), potential execution of arbitrary code via triggering follow-up skills like
execute-unit-tests, and structural changes to the codebase. - Boundary markers: The skill does not provide specific delimiters or "ignore embedded instructions" warnings for the agent to use when processing the code content.
- Sanitization: There is no evidence of sanitization, validation, or filtering of the code content to prevent the interpretation of embedded natural language instructions.
Audit Metadata