The Agent Skills Directory

[SAFE]: The skill includes a dedicated 'Security Protocols' document that mandates a 'Two Repository Strategy' to isolate sensitive data from public code and defines a mandatory 'Three-Check Rule' for git operations to prevent credential leaks.
[EXTERNAL_DOWNLOADS]: TOOLS.md provides installation commands for Bun, UV, and Rustup using curl | bash or curl | sh. These are well-known developer services, and the downloads are documented for initial system setup from trusted domains (bun.sh, astral.sh, and rustup.rs).
[PROMPT_INJECTION]: The skill includes explicit instructions to defend against prompt injection by treating all external content as read-only data and instructs the agent to reject common injection patterns like 'ignore previous instructions'.
[COMMAND_EXECUTION]: The skill implements a 'CLI-First' architecture that maps user intents to deterministic command-line tools. This pattern ensures that operations are predictable, testable, and version-controlled rather than relying on probabilistic agent prompts.

CORE