The Agent Skills Directory

[COMMAND_EXECUTION]: The instrumentation section in SKILL.md instructs the agent to run a local script ./scripts/log-skill.sh to track usage of the skill.
[COMMAND_EXECUTION]: The debugging workflow in workflows/debugging.md recommends using shell commands such as cat .env.local | grep SUPABASE and echo $NEXT_PUBLIC_SUPABASE_URL to verify the local development environment configuration.
[DATA_EXFILTRATION]: The skill includes instructions to read the .env.local file for debugging purposes. While this accesses sensitive configuration data, it is used locally for environment verification and no network exfiltration patterns were identified.
[SAFE]: The provided SQL templates and workflows emphasize security and reliability best practices, such as mandatory Row Level Security (RLS) enforcement, idempotent migration operations using IF NOT EXISTS clauses, and the use of explicit WITH CHECK constraints for write policies.
[SAFE]: The skill uses neutral, instructional language and does not contain any detected prompt injection attempts, obfuscation, or unauthorized persistence mechanisms.

supabase