diagnose
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The reproduction script template (
scripts/hitl-loop.template.sh) serves as an ingestion point where untrusted user input is captured and subsequently processed by the agent. There are no boundary markers or sanitization procedures described to delimit this external content. The skill's broad capabilities for file modification and command execution could be misused if the agent follows malicious instructions embedded in user-provided error data. - [DYNAMIC_EXECUTION]: The skill instructions direct the agent to copy the
hitl-loop.template.shfile, edit it with specific steps, and execute the resulting shell script. This runtime generation and execution of code is a core part of the debugging feedback loop but represents a dynamic execution pattern.
Audit Metadata