agent-bom discover gcp
Inventory GCP AI and MCP Assets to Generate Standardized 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 gcp. It is used for: Inventory GCP AI and MCP Assets to Generate Standardized Lists Full Skill content: https://321skill.com/skills/agent-bom-discover-gcp-x-11/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 of uniformly inventorying AI Agent and MCP-related assets in the GCP cloud environment. In practical operations, teams often need to manually log into the GCP Console or use multiple gcloud commands to query resources such as Vertex AI, Cloud Run, Cloud Functions, and GKE separately, lacking an automated, standardized tool for generating inventory lists. This leads to scattered asset information, making auditing and subsequent analysis difficult.
Usage is straightforward. You only need to install the agent-bom library and configure read-only GCP credentials (ADC, Workload Identity, or short-lived service accounts). Then, a single command can scan a specified project: python examples/operator_pull/gcp_inventory_adapter.py --project $GOOGLE_CLOUD_PROJECT --region us-central1 --source gcp-skill-invoked. It calls the Google Cloud API to discover all AI-related assets and outputs a JSON inventory file compliant with the agent-bom specification to the specified path. The entire process does not require handing long-term credentials to third-party services.
It is ideal for DevOps engineers, agent developers, and project managers who need unified management of GCP AI infrastructure. Particularly for teams that have already deployed multiple AI services (such as Vertex AI models, Cloud Run microservices, Cloud Functions functions), this tool can quickly provide an auditable, scannable asset inventory, supplying foundational data for subsequent compliance checks, cost analysis, and security scanning.
It is recommended to run this tool regularly after each deployment or change to maintain the inventory's real-time status. Note that it only performs discovery operations and will not modify any GCP resources; credentials must have read-only permissions, and using project-level short-lived credentials is advised. The output JSON file can be directly consumed by agent-bom's scanning functionality to achieve closed-loop asset governance.
Key Features
Unlike manually using `gcloud` commands or exporting from the GCP Console, `agent-bom-discover-gcp` automatically identifies AI-related assets such as Vertex AI, Cloud Run, Cloud Functions, and GKE, and generates a standardized `agent-bom` inventory, also supporting MCP-related asset discovery. Furthermore, it does not require users to provide long-term GCP credentials, instead reusing the operator's existing ADC or Workload Identity, offering higher security.
Limitations
Requires Python 3.11+ environment, depends on the `agent-bom` library, supports only the GCP platform, and can only perform read-only discovery operations; it cannot modify or delete resources.
FAQ
What GCP permissions are required?
Project-level read-only permissions are required. It is recommended to use Application Default Credentials, Workload Identity, or short-lived service accounts; do not use long-term service account JSON keys.
How to specify the project to scan?
Specify the project ID via the `--project` parameter, or set the `GOOGLE_CLOUD_PROJECT` environment variable.
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-gcp-x-11/raw/index.md to read the original Skill definition (Markdown format) for agent-bom discover gcp, and install it according to the instructions.
Raw Markdown URL for AI: /skills/agent-bom-discover-gcp-x-11/raw/index.md