skill-forge-quiz
Pass
Audited by Gen Agent Trust Hub on Nov 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest data from external sources, including web search results and user-provided JSON files, which are used to generate quiz content for a database.\n
- Ingestion points: The scripts/import_quiz.js script reads and parses JSON files provided via CLI arguments. SKILL.md instructs the agent to search the web for technical content.\n
- Capability inventory: The skill can write to a local SQLite database (lib/db-writer.js) and spawn local Node.js processes (scripts/import_quiz.js).\n
- Boundary markers: Instructions in SKILL.md provide strict templates and validation steps for the agent to follow when generating content.\n
- Sanitization: The DBWriter uses parameterized SQL queries to prevent SQL injection attacks from the content being imported.\n- [DYNAMIC_EXECUTION]: The skill dynamically resolves and executes a local script to ensure the necessary environment is active.\n
- Evidence: scripts/import_quiz.js uses path.resolve to find a server script in a sibling directory (../../skill-forge-base/lib/server.js) and executes it using spawn if the dashboard port is not responding.\n- [COMMAND_EXECUTION]: The skill instructions and scripts involve executing Node.js commands to perform data operations and manage services.\n
- Evidence: SKILL.md provides commands for importing quizzes (node scripts/import_quiz.js <JSON文件路径>) and scripts/import_quiz.js uses child_process.spawn to run the background dashboard server.
Audit Metadata