The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill fetches markdown documentation and associated code from Snowflake's official GitHub repositories and executes SQL, Python, and Bash code blocks extracted from that content.
[COMMAND_EXECUTION]: The skill uses the Bash and snowflake_sql_execute tools to run commands derived from external Quickstart guides. It also performs git clone operations to retrieve companion repository contents for local execution.
[EXTERNAL_DOWNLOADS]: The skill retrieves content from GitHub APIs and raw content domains (api.github.com, raw.githubusercontent.com) to access documentation and source code.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection through its ingestion of external markdown content.
Ingestion points: External content is fetched in SKILL.md (Step 2) from GitHub or user-provided URLs matching Snowflake's domains.
Boundary markers: The skill does not employ explicit boundary markers or instructions to isolate or ignore embedded instructions within the fetched documentation.
Capability inventory: The skill utilizes snowflake_sql_execute, Bash, Write, and Read tools, providing a broad capability surface if malicious instructions were successfully injected.
Sanitization: No validation or sanitization of the fetched markdown content is performed before processing and execution.

quickstart-guide