The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion and analysis of untrusted external content, creating a vulnerability surface for indirect prompt injection.
Ingestion points: The skill instructs the agent to read SQL migration files under supabase/migrations/, schema dumps such as schema.sql, and project-specific configuration files like CLAUDE.md and AGENTS.md to establish conventions.
Boundary markers: The instructions lack explicit delimiters or "ignore embedded instructions" warnings to prevent the agent from being influenced by natural language commands hidden within the code or documentation being reviewed.
Capability inventory: The skill's capabilities are limited to file system read access and generating a text-based review report. It does not perform network requests, file writes, or command execution based on the ingested data.
Sanitization: No sanitization, escaping, or schema validation is mentioned for the external content before it is processed by the agent.
[SAFE]: The skill incorporates security hardening guidelines, such as requiring SET search_path for SECURITY DEFINER functions to prevent injection attacks and ensuring Row Level Security (RLS) is enabled and correctly configured for all new public-facing tables.

db-reviewer