The Agent Skills Directory

[COMMAND_EXECUTION] (LOW): The skill leverages the Bash and execute_sql tools to run diagnostic commands such as checking connection counts and database bloat. While these are intended for read-only monitoring, the tools themselves provide broad access to the infrastructure and database environment.
[PROMPT_INJECTION] (LOW): The skill is vulnerable to Indirect Prompt Injection (Category 8).
Ingestion points: checklists.md retrieves untrusted data from external sources using railway logs, get_logs, and gh run list.
Boundary markers: Absent. The instructions do not define delimiters or provide specific warnings for the agent to ignore instructions embedded within the logs.
Capability inventory: The agent has access to mcp__plugin_supabase_supabase__execute_sql (allowing database modification) and Bash (environment access) across the entire health check workflow.
Sanitization: Absent. Log and CLI outputs are processed and interpreted without any sanitization or validation of the content.

health-checks