frontend-core
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted user requirements, interaction models, and design assets to generate source code and architectural documentation.
- Ingestion points: Phase 1 and the execution instructions in SKILL.md rely on external user-provided context for classification and requirement gathering.
- Boundary markers: There are no explicit instructions for the agent to use delimiters or ignore potentially malicious instructions embedded in the user's frontend requirements.
- Capability inventory: The skill empowers the agent to generate complete project structures, API integration layers, and configuration files across all referenced phases.
- Sanitization: The skill correctly advises the agent to implement sanitization in the generated output (Phase 11), but lacks internal sanitization logic for its own processing of user input.
Audit Metadata