code-architect
Pass
Audited by Gen Agent Trust Hub on Sep 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides architectural instructions for the agent without any executable code, shell commands, or network operations. The core logic is limited to analyzing project files to generate blueprints.
- [INDIRECT_PROMPT_INJECTION]: The skill processes project files such as CLAUDE.md and AGENTS.md, which could contain instructions intended to influence the agent. 1. Ingestion points: Local codebase files and project guidelines (SKILL.md). 2. Boundary markers: Absent. 3. Capability inventory: Limited to providing textual architecture blueprints and implementation checklists. 4. Sanitization: Absent.
Audit Metadata