NYC

peer-review-simulator

Fail

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: HIGHPROMPT_INJECTIONNO_CODE
Full Analysis
  • Indirect Prompt Injection (HIGH): The skill identifies a workflow to ingest and analyze untrusted external data (manuscripts) using highly privileged tools.
  • Ingestion points: Processes external manuscripts as defined in the 'Usage Guidelines' (Load manuscript, Analyze structure).
  • Boundary markers: Absent; no instructions are provided to the agent to delimit or ignore instructions within the manuscript.
  • Capability inventory: The skill is granted 'Bash' and 'Write' permissions in the 'allowed-tools' section, allowing command execution and file modification.
  • Sanitization: Absent; there is no mention of filtering, escaping, or validating the manuscript content.
  • Risk: An attacker could embed malicious commands in a manuscript that the agent might execute via Bash or Write tools during the analysis phase.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Feb 16, 2026, 08:55 AM