iam-policy-reviewer

Fail

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: HIGHPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [Indirect Prompt Injection] (HIGH): The skill is designed to review IAM policies, which are untrusted data sources that can contain malicious instructions intended to manipulate the agent.
  • Ingestion points: Processes IAM policy JSON/YAML files (implied by the purpose of an iam-policy-reviewer).
  • Boundary markers: Absent. The skill provides no instructions to separate untrusted policy content from agent instructions.
  • Capability inventory: Granted high-privilege tools including Bash, Write, Edit, and Read.
  • Sanitization: None. There is no evidence of sanitization or validation of the input content before processing.
  • [Command Execution] (MEDIUM): The explicit allowance of the Bash tool provides a mechanism for arbitrary code execution on the host system if the agent is successfully influenced by malicious data within the policies it reviews.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Feb 16, 2026, 12:25 PM