anonymize-paper
Pass
Audited by Gen Agent Trust Hub on Jun 30, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill requires the execution of a bundled Python script (
scripts/scan_anonymization.py) to analyze LaTeX documents and directory structures.\n- [EXTERNAL_DOWNLOADS]: The skill instructions direct the agent to fetch venue policies and blind levels from externalcfp_urllinks found in venue profiles or provided by the user.\n- [PROMPT_INJECTION]: The skill is subject to indirect prompt injection because it ingests untrusted LaTeX source files and supplementary content which could contain malicious instructions for the agent.\n - Ingestion points: LaTeX source files (
.tex), bibliography files (.bib), and files within supplementary material directories (scripts/scan_anonymization.py).\n - Boundary markers: No delimiters or explicit instructions are provided to the agent to distinguish between its instructions and the content being analyzed.\n
- Capability inventory: The skill can run local subprocesses (the scanner script), modify files (for anonymization fixes), and perform network operations to fetch policy details.\n
- Sanitization: No sanitization or escaping is performed on the contents of the scanned files before they are processed by the agent.
Audit Metadata