notemdpro-qa-extractor
Pass
Audited by Gen Agent Trust Hub on Apr 16, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted external data within its prompt templates.
- Ingestion points: The
extractOriginalTextfunction reads content from local markdown files (inputPath) using theread_filetool. - Boundary markers: The prompt templates for both Individual and Merged modes use simple text labels like 'Reference Content:' and 'User Question:' without robust delimiters (such as XML tags or unique random separators) to distinguish between instructions and data.
- Capability inventory: The skill has the ability to read and write files (
read_file,write_file) and create directories (mkdir_p). - Sanitization: There is no documented evidence of input sanitization or filtering to prevent embedded instructions in the source text from being interpreted by the LLM.
- [COMMAND_EXECUTION]: The skill documentation describes the execution of shell-equivalent file operations, including
read_file,mkdir_p, andwrite_file. These operations are used to process input documents and generate output files based on user-defined configurations and paths.
Audit Metadata