The Agent Skills Directory

[SAFE]: The skill's primary function is to inventory and analyze local configuration files (agents/, skills/, rules/, and .mcp.json) to estimate token consumption and provide optimization recommendations. These operations are diagnostic and occur within the local project scope.
[DATA_EXPOSURE]: While the skill accesses file content to count tokens and lines, it does not exfiltrate data or access sensitive system paths like credentials or keys. The access is limited to project-specific configuration directories needed for context auditing.
[EXTERNAL_DOWNLOADS]: The instructions reference well-known technology services such as GitHub, Vercel, and Supabase as examples of tools that might be redundantly wrapped by MCP servers. No actual network connections, external downloads, or remote code execution patterns were detected.

context-budget