bedrock-operation-review
Pass
Audited by Gen Agent Trust Hub on Sep 20, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [Indirect Prompt Injection Surface]: The skill processes data from AWS APIs and incorporates it into a report, which presents a surface where untrusted resource metadata could influence the output. * Ingestion points: Resource metadata (names, ARNs, status), CloudWatch metrics, and Service Quota data retrieved via bedrock, cloudwatch, and servicequotas API calls. * Boundary markers: The instructions do not specify explicit delimiters or 'ignore' instructions when interpolating resource data into the markdown report. * Capability inventory: The skill generates a markdown report artifact but does not execute code, write to arbitrary files, or perform network requests. * Sanitization: No explicit sanitization or validation of the ingested resource metadata is described in the instructions.
- [Trusted External References]: The skill provides links to official Amazon Web Services documentation and best practices guides from aws.amazon.com and docs.aws.amazon.com. These references are used to provide context for the review pillars and remediation recommendations.
Audit Metadata