/install awen
GCP Project Bootstrap
Create new Google Cloud projects in a deliberate, least-privilege way.
Workflow
-
Clarify the workload first.
- Cloud Run app
- Vertex AI / Gemini / ML workload
- BigQuery / analytics
- Cloud Storage / file hosting
- Firebase app
- sandbox / experiment / tutorial
-
Collect required inputs before proposing commands.
- project display name
- desired
project_id - org/folder placement if applicable
- billing account availability
- primary region or residency requirement
- services that must work on day one
- whether CI/CD or runtime service accounts are needed
-
Read
references/project-creation-checklist.mdand follow it as the default sequence. -
If billing, API enablement, or service selection is part of the task, read
references/billing-and-apis.md. -
If command examples are needed, read
references/common-gcloud-commands.md.
Operating rules
- Do not assume billing is already linked.
- Do not recommend broad roles like
roles/ownerunless the user explicitly asks and understands the risk. - Prefer least privilege and workload-specific service accounts.
- Call out when a step requires console-only actions, elevated org permissions, or a human with billing admin access.
- If the user has not chosen a region, explain the tradeoff briefly instead of picking randomly.
- If the project is for production, recommend enabling audit-friendly naming and separating runtime/service identities early.
Output expectations
When helping the user, produce:
- a short summary of the intended project purpose
- the exact inputs still missing
- a safe creation sequence
- API list grouped by workload
- IAM/service-account recommendations only if relevant
Common workload branches
Cloud Run
Focus on:
- Artifact Registry
- Cloud Build
- Cloud Run Admin API
- runtime service account
- region choice close to users/data
Vertex AI
Focus on:
- billing must be active
- Vertex AI API
- region support for the intended model/service
- storage buckets for datasets/artifacts if needed
- service account separation for notebooks/jobs/apps
BigQuery
Focus on:
- dataset region planning
- billing confirmation
- BigQuery API
- least-privilege access for analysts vs pipelines
Firebase
Focus on:
- whether this is primarily a Firebase project with GCP underneath
- required console setup beyond raw
gcloud - auth/hosting/firestore specific downstream decisions
- 确保已安装 OpenClaw(本地或 Docker 部署)
- 在对话框中输入安装命令:
/install awen - 安装完成后,直接呼叫该 Skill 的名称或使用
/awen触发 - 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
awen1 是什么?
Create and bootstrap Google Cloud projects for new workloads. Use when the user wants to create a new Google Cloud project, choose a project ID and naming co... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 273 次。
如何安装 awen1?
在 OpenClaw 或 Claude Code 对话框中运行命令「/install awen」即可一键安装,无需额外配置。
awen1 是免费的吗?
是的,awen1 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。
awen1 支持哪些平台?
awen1 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。
谁开发了 awen1?
由 xswazx(@xswazx)开发并维护,当前版本 v1.0.0。