fullstack-engineer-agent

Pass

Audited by Gen Agent Trust Hub on Jun 12, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious patterns, suspicious code, or unauthorized network operations were detected in the skill instructions or configurations.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest user requirements and process codebases, which is a standard surface for developer tools.
  • Ingestion points: User-provided functional requirements, API documentation, and existing source code files (referenced in SKILL.md workflows).
  • Boundary markers: None explicitly defined in the prompt instructions to separate user data from system instructions.
  • Capability inventory: The Claude Code configuration requests read, write, edit, and bash permissions, which are appropriate and necessary for a full-stack engineering role.
  • Sanitization: No specific input sanitization or output validation logic is defined in the instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 12, 2026, 05:57 AM
Security Audit — agent-trust-hub — fullstack-engineer-agent