academic-pipeline
Pass
Audited by Gen Agent Trust Hub on Jul 31, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: Comprehensive analysis of the provided 29 files confirms that the skill functions as an orchestrator for academic workflows without introducing security risks.
- [PROMPT_INJECTION]: No patterns of prompt injection, such as attempts to bypass safety filters or override system instructions, were detected. The 'IRON RULE' labels are used for internal workflow discipline rather than role-play or jailbreak attempts.
- [DATA_EXFILTRATION]: The skill does not access sensitive system files or credentials. Network operations are properly scoped to academic verification tasks, such as querying the Semantic Scholar API and conducting WebSearch for source validation.
- [COMMAND_EXECUTION]: The skill utilizes Tectonic and Pandoc for document compilation and format conversion. These are standard academic utilities, and their usage here is consistent with the skill's stated purpose of finalizing research papers.
- [OBFUSCATION]: No instances of hidden characters, Base64-encoded payloads, or other obfuscation techniques were identified in the instructions or agent configurations.
Audit Metadata