agent-platform-prompt-management

Create, list, retrieve, version, and delete managed prompts in Agent Platform via the Vertex AI SDK.

Updated May 11, 2026
One-click install
npx skills add https://github.com/alon3153/upe-social-publisher --skill agent-platform-prompt-management-alon3153
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: agent-platform-prompt-management
Source: https://github.com/alon3153/upe-social-publisher/tree/main/.agents/skills/agent-platform-prompt-management
Command: npx skills add https://github.com/alon3153/upe-social-publisher --skill agent-platform-prompt-management-alon3153

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires google-cloud-aiplatform, google-genai, and includes references (resource) components.

What problem does it solve? Managing prompts scattered across codebases and consoles is error-prone and hard to govern. This Skill provides structured, safety-tiered operations for the full lifecycle of managed prompts in Agent Platform, so creation, retrieval, versioning, and deletion happen through controlled, confirmed workflows. ## Core Features & Use Cases - Prompt Lifecycle Management: Create, list, retrieve, version, and delete managed prompts using the Vertex AI prompts SDK with Application Default Credentials. - Safety Confirmation Tiers: Read-only operations run immediately, mutating creates require interactive Yes/No confirmation, and destructive deletes require explicit typed confirmation in a separate turn. - Template Variables & Versioning: Assemble prompts with {{variable}} placeholders and retrieve specific version IDs for production prompt governance. - Use Case: A user asks to create a prompt named "Customer Support Greeting" targeting a specific model with a templated greeting; the Skill presents a confirmation card with all parameters, waits for approval, then executes the creation and reports the new prompt ID and version ID. ## Quick Start Ask the assistant to list all managed prompts in your Agent Platform project by providing your Google Cloud project ID and region.

Frequently Asked Questions about agent-platform-prompt-management

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I create a managed prompt in Agent Platform with Python?▼

Build a local Prompt object with prompt_name, prompt_data, and model_name, then call prompts.create_version from the vertexai.preview module. The Skill requires interactive confirmation of the display name, template text, and target model before executing the creation snippet.

How do I list and retrieve prompts using the Vertex AI SDK?▼

Call prompts.list() after vertexai.init with your project and location to enumerate all prompts, and prompts.get(prompt_id=...) to retrieve one. Retrieved prompts expose prompt_id, prompt_data, model_name, and support version_id selection and assemble_contents for variable substitution.

What authentication is required for Agent Platform prompt management?▼

You need Google Cloud Application Default Credentials configured via gcloud auth login and gcloud auth application-default login. The Python environment also needs the google-cloud-aiplatform and google-genai packages installed.

Can I delete a prompt in Agent Platform, and is it reversible?▼

Yes, prompts.delete(prompt_id=...) permanently deletes a prompt using its numeric ID, and the action is irreversible. The Skill enforces explicit typed confirmation in a separate turn before executing any deletion, and re-deleting a missing prompt returns NOT_FOUND treated as success.

Does the Agent Platform prompts SDK support prompt versioning?▼

Yes, prompts.get accepts a version_id parameter to retrieve a specific version, and prompts.create_version creates the prompt and its first version in one call. Best practice is to record version IDs when updating production prompts.

When should I not use this prompt management skill?▼

Do not use it for model training, deploying models to endpoints, or managing prompts outside Agent Platform. It is scoped strictly to managed prompt CRUD and versioning operations through the Vertex AI prompts SDK.