The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes SQL queries through the duckdb command-line tool. In session mode, queries are passed via the shell command duckdb -c "<QUERY>". If the <QUERY> string contains shell metacharacters (e.g., backticks or dollar-sign substitutions) and is not properly escaped by the agent platform, it could lead to arbitrary command execution on the host system.- [INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted data from external files (CSV, Parquet, etc.) and returns the results directly to the agent's context, creating a vulnerability surface where data can influence agent behavior.
Ingestion points: File probing in Step 4 and query execution in Step 5 of SKILL.md.
Boundary markers: Absent; the skill does not provide delimiters or instructions for the agent to ignore instructions embedded within the query results.
Capability inventory: The skill uses the Bash tool to execute duckdb, git, and find commands, with the ability to read and potentially modify the local file system.
Sanitization: No sanitization, filtering, or escaping is applied to the data retrieved from external sources before it is returned to the agent context.

query