cdk-infrastructure
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides instructions for the agent and user to execute standard AWS CDK and Node.js development commands. Evidence includes the use of
npm cifor dependency installation andnpx cdk deployfor provisioning resources (SKILL.md). These are legitimate and necessary for the skill's stated purpose of infrastructure development. - [INDIRECT_PROMPT_INJECTION]: The skill facilitates the generation and modification of complex AWS infrastructure code based on natural language inputs. This creates a risk where malicious instructions could influence the creation of resources with insecure configurations.
- Ingestion points: User prompts requesting infrastructure changes or resource creation (SKILL.md).
- Boundary markers: Absent; there are no specific instructions for the agent to use delimiters or warnings to ignore potentially malicious embedded instructions in user-supplied strings (e.g., resource names, descriptions).
- Capability inventory: The skill enables the creation and modification of sensitive AWS resources including IAM roles (references/iam.md), S3 buckets with custom CORS and lifecycle rules (references/s3.md), and Bedrock AgentCore components like Gateway and Code Interpreter (references/agentcore.md).
- Sanitization: Absent; the reference patterns do not include explicit validation or sanitization logic for user-provided data before it is interpolated into CDK constructs.
Audit Metadata