ljg-paper
Pass
Audited by Gen Agent Trust Hub on Mar 11, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the
dateutility to generate formatted timestamps for filenames and Org-mode headers. While these specific commands (date +%Y%m%dT%H%M%S) are low-risk, they involve subprocess execution. - [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection. It ingest content from untrusted external sources (arXiv, PDFs, and web search results) and processes this data to generate output. An attacker could embed malicious instructions within a paper's text to influence the agent's behavior during analysis.
- Ingestion points: Processes content from
arxiv URL,paper URL,PDF, andWebSearchresults. - Boundary markers: Absent. There are no explicit delimiters or instructions to ignore embedded commands in the source text.
- Capability inventory: File writing to
~/Documents/notes/and shell execution of thedatecommand. - Sanitization: None detected. The skill directly interprets extracted text without filtering for instruction-like patterns.
Audit Metadata