Skills
skills/starchild-ai-agent/official-skills/project-builder/Gen Agent Trust Hub

project-builder

Pass

Audited by Gen Agent Trust Hub on May 19, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the bash tool and shell commands to run, verify, and debug generated code, such as task scripts and backend servers. This is standard functionality for a project engineering tool.
  • [EXTERNAL_DOWNLOADS]: Documentation and code patterns include references to fetching data from established external APIs (e.g., OpenRouter, Twelvedata) and loading visualization libraries from well-known CDNs like jsDelivr and D3js.org.
  • [PROMPT_INJECTION]: The skill addresses the inherent risk of indirect prompt injection in code generation by enforcing a 'Design Gate' in Phase 1. This requires the agent to present a plan and obtain explicit human confirmation before proceeding to write or modify any software.
Audit Metadata
Risk Level
SAFE
Analyzed
May 19, 2026, 12:08 PM
Security Audit — agent-trust-hub — project-builder