alterlab-preprint-deposition
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill makes network requests to well-known academic services, specifically the Sherpa Romeo v2 API (
v2.sherpa.ac.uk) and the bioRxiv Content API (api.biorxiv.org). These connections are used to fetch journal policy information and manuscript publication status, which is consistent with the skill's stated purpose. - [COMMAND_EXECUTION]: The skill uses local Python scripts (
journal_policy.py,preprint_link_check.py, andserver_recommender.py) to perform logic-heavy tasks and API interactions. These scripts are executed via standard platform tools and do not show signs of shell injection or unsafe argument handling. - [INDIRECT_PROMPT_INJECTION]: The skill ingests external data from manuscript titles/abstracts and scholarly API responses. While this represents a potential injection surface, the skill limits the use of this data to submission preparation tasks and lacks dangerous autonomous capabilities that would escalate the risk.
- [DATA_EXFILTRATION]: No access to sensitive system files, environmental variables, or user credentials (such as SSH keys or AWS configs) was found. The use of an API key for Sherpa Romeo is handled via command-line arguments, which is a standard practice for this workflow.
Audit Metadata