skills/aws-samples/sample-well-architected-skills-and-steering/architecture-decision-record/Gen Agent Trust Hub
architecture-decision-record
Pass
Audited by Gen Agent Trust Hub on Jun 16, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill defines a legitimate architectural documentation workflow. It facilitates codebase discovery and design evaluation without executing unauthorized commands, exfiltrating data, or using obfuscated content. The logic is consistent with industry best practices for architecture governance.
- [PROMPT_INJECTION]: The skill interacts with external codebase data (IaC, source code, and configuration manifests) during Step 2. This ingestion of untrusted content represents a potential surface for indirect prompt injection, although the skill's structured template-based output and specific analysis goals mitigate the risk of obeying adversarial instructions embedded in comments or documentation.
- Ingestion points: Local codebase files including CDK, Terraform, application source, and dependency manifests (SKILL.md).
- Boundary markers: Absent; instructions do not explicitly wrap ingested code in delimiters or provide warnings to ignore embedded instructions.
- Capability inventory: File reading capabilities across the codebase and text generation for code, migration scripts, and tests (SKILL.md).
- Sanitization: Codebase content is analyzed and summarized into structured ADR fields without explicit sanitization steps mentioned in the instructions.
Audit Metadata