eb1a-petition
Pass
Audited by Gen Agent Trust Hub on Sep 17, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill fetches official PDF forms (I-140, I-907, G-1145) from the official USCIS website (uscis.gov), which is a trusted government domain.
- [EXTERNAL_DOWNLOADS]: The skill uses the OpenAlex API (api.openalex.org), a well-known academic service, to download research papers and citation data.
- [EXTERNAL_DOWNLOADS]: The skill downloads and installs the
openniwcompanion package from PyPI to provide extended functionality for document harvesting and UI interaction. - [COMMAND_EXECUTION]: The agent is directed to use the
openniwCLI tool to automate tasks such as citation analysis, case auditing, and starting a local web server for user intake. - [INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it ingests untrusted data such as CVs, Scholar profiles, and RFE notices (Ingestion points). The skill utilizes explicit boundary instructions, directing the agent to treat all uploaded documents as data and never as instructions (Boundary markers). The skill possesses significant capabilities including file system access, network operations to fetch papers, and execution of local CLI tools (Capability inventory). The skill implements a form of sanitization by instructing the agent to identify and ignore embedded commands in the processed documents (Sanitization).
Audit Metadata