exam-qa-generator
Pass
Audited by Gen Agent Trust Hub on Jun 25, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to generate and execute local scripts using PowerShell, Bash, Node.js, Bun, or Python to process training material and validate the final JSON output. It specifically suggests creating temporary helper scripts and executing them at runtime.
- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection due to its requirement to ingest and process external content without isolation.
- Ingestion points: The skill reads
CONTENT.mdandSUMMARY.mdfrom a user-provided directory to extract exam metadata and learning content. - Boundary markers: None. The instructions do not define delimiters or provide warnings to the agent to ignore instructions that might be embedded within the training documents.
- Capability inventory: The agent is empowered to execute shell commands, run multiple programming runtimes, and write files to the filesystem.
- Sanitization: No validation or sanitization is performed on the source material before it is processed by the main agent or its subagents.
Audit Metadata