cloud-iam-deep
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill contains an extensive collection of shell commands for AWS, Azure, Google Cloud, and Kubernetes CLI tools. These commands are provided for security professionals to perform cloud resource enumeration and privilege escalation testing.
- [INDIRECT_PROMPT_INJECTION]: The skill documents procedures for extracting identifiers from untrusted external sources, which creates a potential vulnerability surface for indirect prompt injection. 1. Ingestion points: The section on AWS Cognito Identity Pool attacks in SKILL.md instructs on extracting IdentityPoolId values from public JavaScript bundles, mobile app binaries, and source maps. 2. Boundary markers: The instructions do not include boundary markers or warnings to disregard malicious instructions embedded in the extracted data. 3. Capability inventory: The skill leverages powerful cloud management tools including aws, az, gcloud, and kubectl to interact with production environments. 4. Sanitization: The process described lacks validation or sanitization of external input before it is utilized in shell commands.
Audit Metadata