architect
Pass
Audited by Gen Agent Trust Hub on Aug 12, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The methodology requires the agent to generate and execute code for technical spikes and verification tasks. This includes running shell commands and test suites as defined in the stage documents.- [PROMPT_INJECTION]: The skill reads project-related files like PROJECT_STATE.md and build plans to maintain context. This creates a surface for indirect injection if these files contain untrusted data. (Ingestion points: Project metadata and plan files. Boundary markers: Absent. Capability inventory: File system writes and command execution. Sanitization: Absent).
Audit Metadata