s3-bucket-policy-generator

Fail

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: HIGHCOMMAND_EXECUTIONPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
  • [COMMAND_EXECUTION] (HIGH): The skill metadata requests Bash(aws:*) tool access, which allows for the execution of arbitrary AWS CLI commands. This permission level is excessive for policy generation and creates a massive attack surface.
  • [PROMPT_INJECTION] (HIGH): The skill has a high-risk Category 8 (Indirect Prompt Injection) vulnerability surface. Ingestion points: User-provided requests triggered by 's3 bucket policy generator'. Boundary markers: None. Capability inventory: High (Bash(aws:*), Read, Write, Edit). Sanitization: None. Attackers could inject malicious instructions that leverage these high-privilege tools.
  • [DATA_EXFILTRATION] (HIGH): The combination of broad shell access and file read capabilities allows the agent to access sensitive configuration files and potentially exfiltrate AWS credentials or account data if prompted maliciously.
Recommendations
  • AI detected serious security threats
Audit Metadata
Risk Level
HIGH
Analyzed
Feb 16, 2026, 08:01 AM