The Agent Skills Directory

[SAFE]: The skill focuses on operational safety for Roblox development toolchains (Rojo, Argon, etc.). It emphasizes "do not destroy work" and requires the agent to confirm the "source of truth" before performing any file operations.
[SAFE]: Instructions explicitly mandate that the agent should never start, stop, or reconfigure sync sessions unasked, and must run preflight checks before any writes.
[SAFE]: External references point to relative paths within a shared pool (../best-practices/references/), which is standard for modular skill sets. No external network downloads or remote code executions are present.
[SAFE]: The trigger queries in evals/trigger-queries.json correctly distinguish between operational tooling tasks (which trigger this skill) and Luau coding/diagnostic tasks (which are handed off to other skills).

studio-ops