backward-traceability
Warn
Audited by Gen Agent Trust Hub on Apr 21, 2026
Risk Level: MEDIUMREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
- [DYNAMIC_EXECUTION]: The reference file 'references/traceability-patterns.md' describes an implementation of the '\num' command using Python's 'eval()' function to process strings extracted from LaTeX documents. This documentation promotes a high-risk pattern that allows arbitrary code execution if the LaTeX source is maliciously crafted or contains untrusted input.
- [INDIRECT_PROMPT_INJECTION]: The Python script 'scripts/ref_numeric_values.py' is designed to ingest and process data from external LaTeX files provided as an argument.
- Ingestion points: The script reads the content of 'tex_file' and 'code_output' paths provided via CLI arguments.
- Boundary markers: No delimiters or boundary warnings are present in the script's processing logic.
- Capability inventory: The script performs file reads and regex-based scanning; it does not contain high-risk capabilities such as network exfiltration or shell command execution.
- Sanitization: No sanitization or validation of the LaTeX content is performed before regex extraction.
Audit Metadata