707-technologies-hexagonal-architecture
Pass
Audited by Gen Agent Trust Hub on Sep 26, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill mandates the execution of build and verification commands to ensure workspace integrity during architectural changes.
- Evidence: Instructions in
SKILL.mdandreferences/707-technologies-hexagonal-architecture.mdrequire running./mvnw compile,mvn compile,./mvnw clean verify, andmvn clean verifybefore and after modifications. - Context: These are standard development tasks for Java projects and are essential for the skill's primary purpose of architectural remediation.
- [INDIRECT_PROMPT_INJECTION]: The skill analyzes external data (Java source code and Maven configurations) to provide guidance, which presents a surface for indirect prompt injection.
- Ingestion points: The skill reads packages, modules, imports, Maven modules, and tests (found in
SKILL.mdWorkflow step 1). - Boundary markers: No explicit delimiters or instructions to ignore potential prompts embedded in source code comments or metadata are present.
- Capability inventory: The skill has the capability to execute shell commands (via Maven) and modify local source files to remediate architecture violations.
- Sanitization: No specific sanitization or filtering logic for ingested source code is described in the skill instructions.
Audit Metadata