research-paper-writer
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- [Prompt Injection] (SAFE): No malicious instructions, bypass attempts, or extraction patterns were found within the instructional content.
- [Data Exposure & Exfiltration] (SAFE): The skill lacks network access and does not reference any sensitive local file paths.
- [Unverifiable Dependencies & Remote Code Execution] (SAFE): There are no remote script executions or external package dependencies listed in the configuration files.
- [Command Execution] (SAFE): The provided JavaScript file is a non-functional stub that performs no system commands or subprocess spawning.
- [Indirect Prompt Injection] (SAFE): Evidence Chain: 1. Ingestion points: User-provided research data and topics (SKILL.md Section 1). 2. Boundary markers: Not present. 3. Capability inventory: No network, shell, or file system access. 4. Sanitization: Not present. Despite the lack of sanitization, the skill is safe because it lacks any capabilities that could be exploited by malicious data.
Audit Metadata