Conference Paper Writer
Pass
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: LOWNO_CODE
Full Analysis
- [NO_CODE] (SAFE): The skill consists entirely of descriptive text, tables, and checklists. There are no associated scripts (.sh, .py, .js), binary executables, or configuration files that define logic for the agent to execute.
- [DATA_EXPOSURE] (SAFE): The document does not contain any hardcoded credentials, sensitive file paths, or network communication commands (e.g., curl, wget).
- [PROMPT_INJECTION] (SAFE): No instructions designed to override the agent's safety protocols or system prompts were detected. The language is purely instructional and follows a standard workflow format.
- [INDIRECT_PROMPT_INJECTION] (INFO): The workflow involves 'Discovery' and 'Research' phases that ingest external requirements. However, because the skill has no capability to write to the filesystem, execute shell commands, or make network requests, there is no technical surface for an attacker to leverage via data ingestion.
Audit Metadata