skills/xiaolai/vmark/mcp-dev/Gen Agent Trust Hub

mcp-dev

Pass

Audited by Gen Agent Trust Hub on Apr 27, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill includes a script 'scripts/scan_mcp.sh' which executes 'rg' (ripgrep) to find occurrences of 'mcp' in specified local directories. This is a standard utility for code discovery and does not perform any dangerous operations.
  • [DATA_EXFILTRATION]: The skill references local configuration files like '.mcp.json' and '.claude/settings.local.json' for development purposes. No network tools are used and no instructions for exfiltration are present.
  • [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection because it reads external project documentation. 1. Ingestion points: 'dev-docs/' and 'src/' directories. 2. Boundary markers: Absent. 3. Capability inventory: Updating local '.mcp.json' configuration files. 4. Sanitization: Absent. This surface is expected for development-oriented tools and is considered low risk in this context.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 27, 2026, 01:38 PM