database-migration-service-expertise
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- Indirect Prompt Injection Surface: The skill is designed to ingest and analyze untrusted data, such as CloudWatch logs and database task statistics, which could potentially contain malicious instructions intended to influence the agent's behavior.
- Ingestion points: CloudWatch Logs and DMS table statistics (SKILL.md, Step 3 and 4).
- Boundary markers: Absent; the instructions do not specify the use of delimiters or specific instructions to ignore embedded commands within the log data.
- Capability inventory: The skill uses AWS CLI commands (
aws dms,aws cloudwatch,aws logs) to interact with the cloud environment. - Sanitization: Absent; while the skill uses filter patterns for logs, the resulting content is directly analyzed by the agent without explicit sanitization steps.
- Command Execution: The skill instructs the agent to perform discovery and data collection using the AWS CLI. While primarily read-only, it includes capabilities such as
dms:TestConnectionwhich involves active network interaction with database endpoints. - External Resource Reference: The README.md file directs users to a CloudFormation template hosted in an official AWS GitHub repository for configuring necessary IAM permissions.
Audit Metadata