The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes external data from a database and provides instructions for presenting it without using boundary markers.
Ingestion points: The skill ingests untrusted data from the results of the sql_db_query tool.
Boundary markers: There are no instructions to use delimiters or warnings to separate database content from the agent's instructions.
Capability inventory: The skill has database read capabilities through sql_db_query and task planning via write_todos.
Sanitization: No input validation or escaping mechanisms are defined for the data retrieved from the database.
[SAFE]: The skill implements a significant security control by explicitly prohibiting the use of Data Manipulation Language (DML) commands (INSERT, UPDATE, DELETE, DROP), which ensures the agent can only perform read-only operations.
[SAFE]: Quality guidelines enforce the use of LIMIT on all queries, which protects against potential resource exhaustion from excessively large database responses.

query-writing