The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill includes instructions to retrieve and decode Kubernetes secrets in plaintext (kubectl get secret supabase-secrets -o jsonpath='{.data.api-key}' | base64 -d), which exposes sensitive credentials to the console and command logs.
[COMMAND_EXECUTION]: The runbook provides commands for high-impact database administrative actions, such as pg_cancel_backend and pg_terminate_backend, which can disrupt active database connections and should be used with caution.
[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes external data from the Supabase status API and database query results, creating a surface for indirect prompt injection if the external data source is compromised.
Ingestion points: API responses from status.supabase.com and results from database queries targeting pg_stat_activity and pg_policy.
Boundary markers: The skill does not employ explicit delimiters or instructions to ignore potential commands embedded within external data.
Capability inventory: The skill utilizes significant capabilities including shell execution (Bash), database querying (psql), and file system reading (Read).
Sanitization: No explicit sanitization or validation logic is present for the data retrieved from external sources before it is processed by the agent.

supabase-incident-runbook