← Back to Skills Marketplace
1095
Downloads
2
Stars
3
Active Installs
2
Versions
Install in OpenClaw
/install infrastructure
Description
Design, provision, and connect cloud resources across servers, networks, and services.
README (SKILL.md)
Scope
This skill:
- ✅ Guides architecture decisions
- ✅ Provides provisioning commands for user to run
- ✅ Documents infrastructure patterns
User-driven model:
- User provides cloud credentials when needed
- User runs provisioning commands
- Skill guides decisions and generates commands
This skill does NOT:
- ❌ Store or access cloud credentials directly
- ❌ Run provisioning commands automatically
- ❌ Modify infrastructure without user confirmation
For implementation: User runs commands skill provides, or uses server skill for execution.
Quick Reference
| Topic | File |
|---|---|
| Architecture patterns | patterns.md |
| Provider commands | providers.md |
| Backup strategies | backups.md |
Core Rules
1. User Runs Commands
Skill generates commands, user executes:
Agent: "To create the server, run:
hcloud server create --name web1 --type cx21 --image ubuntu-24.04
This requires HCLOUD_TOKEN in your environment."
User: [runs command]
2. Required Tools (User Installs)
| Provider | Tool | Install |
|---|---|---|
| Hetzner | hcloud |
brew install hcloud |
| AWS | aws |
brew install awscli |
| DigitalOcean | doctl |
brew install doctl |
| Docker | docker |
Docker Desktop |
3. Credential Handling
- User sets credentials in their environment
- Skill never stores or logs credential values
- Commands reference env vars:
$HCLOUD_TOKEN,$AWS_ACCESS_KEY_ID
4. Architecture Guidance
| Stage | Recommended |
|---|---|
| MVP | Single VPS + Docker Compose |
| Growth | Dedicated DB + load balancer |
| Scale | Multi-region + CDN |
5. Decision Framework
| Question | Answer |
|---|---|
| How to structure infra? | ✅ This skill |
| Should I add another server? | ✅ This skill |
| How to configure nginx? | Use server skill |
| How to write Dockerfile? | Use docker skill |
6. Backup Strategy
| Data | Method | Frequency |
|---|---|---|
| Database | pg_dump → S3/B2 | Daily |
| Volumes | Snapshots | Weekly |
| Config | Git | Every change |
Usage Guidance
This skill is essentially documentation and examples — not executable code pushed by the skill — so review and adapt everything before running. Specific suggestions: 1) Inspect backup scripts (pg_dump, aws s3 cp and cleanup logic) and replace placeholder bucket names; validate the aws s3 ls parsing/cleanup logic before enabling retention deletion. 2) Use least-privileged IAM credentials and short-lived tokens where possible; never paste long-lived root credentials into chats. 3) Test backup and restore procedures in a safe environment before relying on them. 4) When installing cron jobs or scripts into /opt or /etc, check file permissions and ownership. 5) Verify CLI tool installation and versions from official sources. If you want stronger metadata safety, ask the skill author to declare required env vars and binaries in the registry so automated checks can surface mismatches.
Capability Analysis
Type: OpenClaw Skill
Name: infrastructure
Version: 1.0.1
The skill bundle is benign. The `SKILL.md` clearly outlines a user-driven model, stating the agent does not store credentials, run commands automatically, or modify infrastructure without user confirmation. The `backups.md` and `providers.md` files contain shell script examples for legitimate infrastructure tasks like creating servers, managing firewalls, and performing database backups (including cleanup with `aws s3 rm` for retention), all intended for user execution. There is no evidence of prompt injection, data exfiltration to unauthorized endpoints, or other malicious intent; the commands are standard for cloud management and align with the skill's stated purpose.
Capability Assessment
Purpose & Capability
The name/description match the provided content: architecture patterns, provider CLI commands, and backup scripts. The skill is instruction-only and intends the user to run commands. It recommends common provider CLIs (hcloud, aws, doctl, docker) which are reasonable for the stated purpose. Minor documentation/metadata mismatch: registry metadata lists no required binaries/env vars while SKILL.md documents required tools and references environment variables the user must set.
Instruction Scope
SKILL.md and the companion files show only guidance, example CLI invocations, and example scripts (cron entry, backup script) that assume the user provides credentials/environment variables and executes commands locally. There are no hidden endpoints, obfuscated code, or instructions to exfiltrate data. The scripts do reference env vars (e.g., $DATABASE_URL, $HCLOUD_TOKEN, AWS creds) and write outputs to /tmp and suggested cron paths — these are expected for backup/infra tooling but should be reviewed before execution.
Install Mechanism
No install spec or remote downloads are present; all files are documentation and sample scripts. The skill only recommends commonly used community CLIs (brew install hcloud/awscli/doctl) and Docker Desktop — nothing fetched or installed automatically by the skill itself.
Credentials
The skill does not declare required env vars in registry metadata, but its content repeatedly references credentials and connection strings (HCLOUD_TOKEN, AWS_ACCESS_KEY_ID, $DATABASE_URL, etc.). This is consistent with its 'user-driven credentials' model, but users should be aware they must supply appropriate credentials locally. Use least-privilege keys and avoid pasting sensitive tokens into shared/public contexts.
Persistence & Privilege
The skill is not always-enabled and is user-invocable. It does not request persistent installation or modify other skills or system-wide settings. It does provide example scripts that a user might install (cron job, /opt/scripts), but these are user actions rather than automatic behavior by the skill.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install infrastructure - After installation, invoke the skill by name or use
/infrastructure - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.1
User-driven credential model, explicit tool requirements
v1.0.0
Initial release
Metadata
Frequently Asked Questions
What is Infrastructure?
Design, provision, and connect cloud resources across servers, networks, and services. It is an AI Agent Skill for Claude Code / OpenClaw, with 1095 downloads so far.
How do I install Infrastructure?
Run "/install infrastructure" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Infrastructure free?
Yes, Infrastructure is completely free (open-source). You can download, install and use it at no cost.
Which platforms does Infrastructure support?
Infrastructure is cross-platform and runs anywhere OpenClaw / Claude Code is available (linux, darwin, win32).
Who created Infrastructure?
It is built and maintained by Iván (@ivangdavila); the current version is v1.0.1.
More Skills