The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructions and scripts suggest the use of standard development commands like npm run dev, npm run build, and npx create-mcp-use-app. These are expected for a project scaffolding and development workflow.\n- [EXTERNAL_DOWNLOADS]: The skill references various official documentation sites and GitHub repositories for research purposes, including sources from OpenAI, Anthropic, and Vercel. These are well-known and trusted services.\n- [DATA_EXFILTRATION]: The instructions and scripts emphasize security best practices, specifically warning against the inclusion of secrets in tool results, metadata, or logs.\n- [REMOTE_CODE_EXECUTION]: The provided Python scripts (bootstrap_harness.py and check_project.py) operate on local directories to manage project files and do not perform any remote code execution or unsafe dynamic code loading.\n- [PROMPT_INJECTION]: The skill includes safety-focused guidelines and evaluation prompts designed to test and prevent prompt injection in the resulting MCP applications.

mcpeas