paper-image-extractor
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill downloads paper source packages from arXiv's official e-print service (
arxiv.org). This is a well-known academic repository used as intended for research workflows. - [COMMAND_EXECUTION]: The skill uses the
run_terminaltool to execute a bundled Python script for processing PDF and source files. The script is used locally and does not fetch or execute arbitrary remote scripts. - [INDIRECT_PROMPT_INJECTION]: The skill has a potential attack surface where user-provided paper IDs are processed by scripts and shell commands.
- Ingestion points: The user-supplied
[PAPER_ID]inSKILL.mdandscripts/extract_images.py. - Boundary markers: Absent in the shell command templates in
SKILL.md. - Capability inventory: Terminal execution (
run_terminal), file writes (write_file), and network access viarequestsorurllib. - Sanitization: The
extract_images.pyscript includes a robust path traversal check when extracting source files from tarballs (if member.name.startswith('/') or '..' in member.name: continue), preventing 'Zip Slip' style attacks.
Audit Metadata