agent-bom discover snowflake
Inventorying Snowflake AI Assets to Generate Standardized Inventory Lists
Install & Use
Copy this prompt and send it to your AI assistant (Claude / Cursor / TRAE / Codex / WorkBuddy etc.) to auto-install:
Help me install this AI Skill: agent-bom discover snowflake. It is used for: Inventorying Snowflake AI Assets to Generate Standardized Inventory Lists Full Skill content: https://321skill.com/skills/agent-bom-discover-snowflake-x-3/raw/index.md Read that page and install it.
The prompt includes a link to the full Skill content. You can also view the full content.
This Skill addresses the challenge enterprises face in systematically inventorying AI assets within Snowflake. In real-world operations, a Snowflake environment may contain various AI-related resources such as Cortex AI, Snowpark, Notebooks, Streamlit, and MCP. Manual inventorying is prone to omissions and lacks a unified format, which hinders subsequent auditing and asset management.
Usage is straightforward. You only need read-only credentials for Snowflake (SSO, OAuth, or key-pair authentication is recommended) and ensure a Python 3.11+ environment with the agent-bom Snowflake extension installed. After running the discovery script, it automatically queries your account for Cortex, Snowpark, Notebook, Streamlit, MCP, and AI observability assets. It then generates a JSON inventory file compliant with the agent-bom specification and writes it to your specified path. The entire process does not require handing over long-term credentials to agent-bom; credentials remain solely within the operator's environment.
It is well-suited for operations engineers, data analysts, and agent developers who need to manage Snowflake AI infrastructure. Teams already deploying Cortex AI, Streamlit applications, or Snowpark pipelines can particularly benefit, using this tool to quickly obtain a standardized asset inventory for compliance audits, cost analysis, or migration planning.
We recommend using this tool uniformly for initial inventorying or periodic audits. Note that it only supports read-only operations and will not modify any Snowflake resources. It also requires the operator to have read-only role permissions for the corresponding account. Prioritize using key-pair or OAuth authentication to avoid transmitting passwords or tokens in conversations.
Key Features
Unlike Snowflake's native INFORMATION_SCHEMA queries, this Skill specifically discovers AI-related assets (Cortex, Streamlit, MCP, etc.) and outputs a standardized `agent-bom` JSON format, facilitating subsequent scanning and integration. Additionally, it does not store long-term credentials, offering higher security.
Limitations
Requires a Python 3.11+ environment, depends on the `agent-bom` Snowflake extension, supports only read-only queries, and cannot discover non-AI traditional database objects.
FAQ
What Snowflake permissions are required?
A read-only role for the account is required. It's recommended to use a read-only role authorized by ACCOUNTADMIN or SECURITYADMIN to ensure the ability to query metadata for Cortex, Streamlit, etc.
Which authentication methods are supported?
SSO, OAuth, and key-pair authentication are prioritized; password usage should be avoided. Configuration can be done via the SNOWFLAKE_PRIVATE_KEY_PATH or SNOWFLAKE_AUTHENTICATOR environment variables.
Installation guide for AI assistants
If your AI coding assistant (Claude Code, Cursor, TRAE etc.) can see this page, send it this message to auto-install:
Visit https://321skill.com/skills/agent-bom-discover-snowflake-x-3/raw/index.md to read the original Skill definition (Markdown format) for agent-bom discover snowflake, and install it according to the instructions.
Raw Markdown URL for AI: /skills/agent-bom-discover-snowflake-x-3/raw/index.md