frontend-developer

Pass

Audited by Gen Agent Trust Hub on May 4, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No suspicious behavior or security vulnerabilities detected. The skill provides a structured workflow for senior frontend development tasks.\n- [PROMPT_INJECTION]: The instructions do not contain attempts to override agent behavior, bypass safety constraints, or extract system prompts.\n- [DATA_EXFILTRATION]: No network operations, hardcoded credentials, or access to sensitive local files (e.g., .ssh, .aws) were found.\n- [REMOTE_CODE_EXECUTION]: There are no patterns involving the download or execution of remote scripts or unpinned dependencies.\n- [INDIRECT_PROMPT_INJECTION]: While the skill ingests external project context, it lacks the tools or capabilities (e.g., shell access, network ops) required to exploit such an injection. \n
  • Ingestion points: Project context received from the context-manager via assets/context_request.json.\n
  • Boundary markers: Absent.\n
  • Capability inventory: Limited to text/code generation; no tool-based capabilities listed in allowed-tools.\n
  • Sanitization: Absent.
Audit Metadata
Risk Level
SAFE
Analyzed
May 4, 2026, 07:12 AM
Security Audit — agent-trust-hub — frontend-developer