aws-solution-architect
Pass
Audited by Gen Agent Trust Hub on Nov 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted JSON data from external files to generate architecture recommendations and infrastructure templates.
- Ingestion points: The scripts
architecture_designer.py,cost_optimizer.py, andserverless_stack.pyread application requirements and resource inventories from JSON files provided by the user. - Boundary markers: No explicit delimiters or instructions are used to separate user-provided data from the agent's internal logic during processing.
- Capability inventory: The skill includes file-writing capabilities and instructions for the agent to execute infrastructure deployment commands.
- Sanitization: Basic regex validation is performed on the
app_nameandregionfields inserverless_stack.py, but other data fields are processed and interpolated without extensive sanitization. - [DYNAMIC_EXECUTION]: The
scripts/serverless_stack.pyscript generates infrastructure-as-code (IaC) templates at runtime. - Evidence: The script dynamically assembles CloudFormation, CDK, and Terraform configurations using string interpolation based on user-supplied parameters. This represents the dynamic creation of executable infrastructure code.
- [COMMAND_EXECUTION]: The skill instructions direct the agent to perform multiple shell-based operations.
- Evidence:
SKILL.mdandreferences/workflow-and-usage.mdguide the agent to execute local Python scripts and deploy generated templates using command-line tools such assam,cdk, andterraform. Additionally,serverless_stack.pyincludes a parameter to write its output to an arbitrary local file path.
Audit Metadata