菜单
Orchestrates setup, testing, and teardown of the jira-mcp-direct recipe under agy-recipes/. Use when the user requests deployment, replication, or testing of the two direct Atlassian Jira MCP connectors (Rovo hosted + Custom Cloud Run).
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Expert guide for building, deploying, and debugging Google ADK (Agent Development Kit) agents. Use for creating LLM agents, workflows, tools, and multi-agent systems.
Enforces strict guidelines for creating, modifying, and refactoring reproducible recipes under the `agy-recipes/` folder. Use when developing setup/teardown scripts, IAM configurations, API enablement, or workflows for replicating deployments.
Replicates, configures, and spins up the local Gemini Enterprise MCP Co-work Portal web application. Use when the user requests replicating, deploying, or testing the co-work portal web app.
Sets up, tests, and tears down a GCS Data Store linked to a Gemini Enterprise engine, with support for Workforce Identity Federation (WIF) access. Use when the user requests linking Cloud Storage buckets to Gemini Enterprise or testing WIF-based GCS search.
Orchestrates the replication, environment provisioning, container builds, and deployment of the Shutter Vibe Engine on Google Cloud. Use when the user requests deployment, replication, or troubleshooting of the multimodal vibe search system.
Guides the creation of Google ADK (Agent Development Kit) agents. Use when the user asks to create, configure, or understand ADK agents, workflows, or tools.
| name | replicating-jira-mcp-direct |
| description | Orchestrates setup, testing, and teardown of the jira-mcp-direct recipe under agy-recipes/. Use when the user requests deployment, replication, or testing of the two direct Atlassian Jira MCP connectors (Rovo hosted + Custom Cloud Run). |
This skill guides the agent and user through deploying, validating, and destroying the direct Atlassian Jira MCP integration connectors in Gemini Enterprise.
.env file exists in agy-recipes/jira-mcp-direct/ with:
GE_PROJECT_ID (GCP Project ID, e.g. vtxdemos)GE_PROJECT_NUMBER (GCP Project Number, e.g. 254356041555)GE_ENGINE_ID (Gemini Enterprise Engine ID, e.g. jira-testing_1778158449701)ATLASSIAN_CLIENT_ID (Standard Atlassian App Client ID, for Option C)ATLASSIAN_CLIENT_SECRET (Standard Atlassian App Client Secret, for Option C)gcloud credentials authorized for your target project.Execute the setup script:
uv run agy-recipes/jira-mcp-direct/scripts/setup.py
Because Atlassian requires interactive user authorization (3LO), you must complete the final wiring in the Cloud Console:
Atlassian Hosted Rovo Connector (jiramcp-rovo-recipe):
jiramcp-rovo-recipe_mcp_data -> Actions tab.searchJiraIssuesUsingJql, getJiraIssue).~/.secrets/atlassian-rovo-dcr-ge.json.Custom Cloud Run MCP Connector (jiramcp-custom-recipe):
jiramcp-custom-recipe_mcp_data -> Actions tab.Once both connectors are active (showing green checks in the console), run the test script:
uv run agy-recipes/jira-mcp-direct/scripts/test_recipe.py
To delete all resources created by this recipe:
uv run agy-recipes/jira-mcp-direct/scripts/teardown.py
This cleanly removes both datastores in Gemini Enterprise, detaches them from the engine, and deletes the Cloud Run server.