capacitor-mcp

Pass

Audited by Gen Agent Trust Hub on Sep 10, 2026

Risk Level: SAFECOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill includes explicit shell commands and configuration blocks for registering the 'capacitor' MCP server in environments like Claude Code, Cursor, and VS Code.
  • [REMOTE_CODE_EXECUTION]: For clients requiring local command execution, the skill recommends running the @capawesome/capacitor-mcp package via npx. This package is a vendor-provided tool used to proxy requests to the hosted MCP server.
  • [INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface through its reliance on local project metadata.
  • Ingestion points: Retrieves the @capacitor/core version from the project's package.json file (SKILL.md).
  • Boundary markers: None are defined for the version parameter used in tool invocations.
  • Capability inventory: Uses documentation content fetched from the remote server to guide agent behavior (SKILL.md).
  • Sanitization: No validation or sanitization of the project version or documentation content is performed.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 10, 2026, 02:00 PM
Security Audit — agent-trust-hub — capacitor-mcp