mcp-server-integration
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill correctly instructs the use of GitHub Secrets for managing sensitive tokens, avoiding hardcoded credentials.
- [EXTERNAL_DOWNLOADS]: The skill references and installs official Model Context Protocol (MCP) and Playwright packages from the public NPM registry, which is a well-known and reputable service.
- [COMMAND_EXECUTION]: Provides configuration for running local MCP server executables and browser automation tools such as npx -y @playwright/mcp@latest.
- [INDIRECT_PROMPT_INJECTION]: The skill provides tools for ingesting data from external sources such as the filesystem and web browsing via Playwright, creating a potential attack surface for indirect prompt injection. 1. Ingestion points: filesystem-read_text_file and playwright-browser_navigate in SKILL.md. 2. Boundary markers: None explicitly mentioned in the configuration examples. 3. Capability inventory: filesystem-write_file, github-create_issue, and shell in SKILL.md. 4. Sanitization: Not explicitly implemented in the provided patterns.
Audit Metadata