The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill provides the query_render_postgres tool which accepts and executes SQL queries provided by the user. This creates a surface where the agent could be manipulated into executing queries that disclose sensitive data or interact with attacker-controlled content stored in the database.
Ingestion points: SKILL.md (the sql argument for the query_render_postgres tool).
Boundary markers: Absent; there are no instructions or delimiters provided to ensure the agent distinguishes between its instructions and the user-provided SQL.
Capability inventory: Includes query_render_postgres (read-only database access), get_postgres, list_postgres_instances, and get_metrics.
Sanitization: Absent; the skill instructions do not include logic for sanitizing or validating the SQL string before execution.
[SAFE]: The connection strings found in references/connection-guide.md (e.g., postgres://user:password@hostname/dbname) are generic templates used for educational purposes and do not contain hardcoded credentials or actual secrets.

render-postgres