academic-pipeline

Pass

Audited by Gen Agent Trust Hub on Aug 16, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill references the execution of local scripts and standard developer tools for auditing and testing. It provides examples for running unit tests via python3 -m unittest and utility scripts such as scripts/run_codex_audit.sh and scripts/check_literature_corpus_schema.py.
  • [COMMAND_EXECUTION]: External command-line tools pandoc and tectonic are used for document conversion and LaTeX compilation to produce academic manuscripts.
  • [EXTERNAL_DOWNLOADS]: The skill utilizes the Semantic Scholar API and general WebSearch queries for research verification, source tracing, and fact-checking.
  • [PROMPT_INJECTION]: The skill processes untrusted external data such as research drafts and reviewer feedback. It addresses this risk through a mandatory integrity_verification_agent that implements multi-phase verification of citations, data, and claims.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 16, 2026, 05:26 AM
Security Audit — agent-trust-hub — academic-pipeline