scholar-evaluation
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill communicates with the OpenRouter API (openrouter.ai) to facilitate AI-powered scientific schematic generation and quality review. This is a well-known service and the usage is consistent with the skill's primary functionality.
- [COMMAND_EXECUTION]: The script
scripts/generate_schematic.pyusessubprocess.runto execute a local companion script (scripts/generate_schematic_ai.py). This is a legitimate implementation of the skill's diagram refinement logic. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted scholarly content which represents an indirect injection surface. While this is inherent to the task of research evaluation, the skill currently lacks explicit boundary delimiters to isolate external text from internal instructions.
- Ingestion points: Research papers, proposals, and literature reviews processed in the
scholar-evaluationworkflow. - Boundary markers: No explicit markers or "ignore" instructions found in SKILL.md.
- Capability inventory: Local file writing for figures and reports; network access to the OpenRouter API.
- Sanitization: None detected for processed scholarly text.
Audit Metadata