ieee-mg-writing
Pass
Audited by Gen Agent Trust Hub on Sep 7, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill package is composed entirely of static Markdown templates and YAML configuration files. It does not include any executable scripts, binary files, or tool definitions, effectively eliminating the risk of local or remote code execution.
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided technical information to generate academic text, creating a surface for indirect prompt injection.
- Ingestion points: User requests containing technical topics, simulation data, and research contributions as described in the routing logic of
SKILL.md. - Boundary markers: The skill includes a "Judging Boundaries" section in
SKILL.mdthat defines explicit rules for when the agent must query the user and when to stop generation to prevent hallucination. - Capability inventory: The skill has no enabled tools, network capabilities, or file-system writing permissions.
- Sanitization: The skill relies on the primary agent's underlying safety filters for processing external technical content.
Audit Metadata