convert-web-app

Pass

Audited by Gen Agent Trust Hub on Apr 1, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill fetches reference code and examples from the official Model Context Protocol GitHub repository (https://github.com/modelcontextprotocol/ext-apps.git) to provide context for the agent and user.
  • [COMMAND_EXECUTION]: The skill instructs the agent to use standard development commands, such as git clone for documentation and npm install for project dependencies (@modelcontextprotocol/ext-apps, @modelcontextprotocol/sdk, etc.).
  • [PROMPT_INJECTION]: The instructions focus on technical implementation and architectural guidance for a specific protocol. No patterns for bypassing safety filters, overriding system instructions, or extracting system prompts were detected.
  • [DATA_EXFILTRATION]: No hardcoded credentials, sensitive file access patterns (e.g., accessing .ssh or .env files), or unauthorized network exfiltration attempts were found. The skill correctly identifies that localStorage is not available in sandboxed environments, which is a security best practice.
  • [INDIRECT_PROMPT_INJECTION]: The skill involves an attack surface where the agent analyzes existing web application code which could contain untrusted data.
  • Ingestion points: Step 1 requires the agent to analyze the user's existing web application files.
  • Boundary markers: No specific delimiters or warnings to ignore embedded instructions in the analyzed code are provided.
  • Capability inventory: The skill involves package installation (npm install), repository cloning (git clone), and build processes (vite build).
  • Sanitization: No explicit sanitization or validation of the analyzed application content is mentioned.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 1, 2026, 02:34 PM