The Agent Skills Directory

[SAFE]: No malicious patterns, obfuscation, or unauthorized data exfiltration attempts were detected in the skill instructions or provided code examples.
[DATA_EXPOSURE]: The skill is designed to discover and preview database content. It correctly recommends using process.env.MOTHERDUCK_TOKEN for authentication, which is a standard security best practice for managing secrets.
[INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface because it processes data from external sources (MotherDuck database catalogs and table content).
Ingestion points: Data entering the context via catalog queries (duckdb_tables, duckdb_columns) and row previews.
Boundary markers: None identified in the provided templates.
Capability inventory: The skill uses SQL query tools (query, query_rw) to interact with the database.
Sanitization: No specific sanitization or filtering of database-returned strings is mentioned. While this is a common risk for data-exploration skills, users should be aware that the agent could theoretically be influenced by malicious content stored within database metadata or table rows.

motherduck-explore