More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    OpenStack Cloud Operations

    1

    Give your AI agent structured OpenStack investigation and resource-operation workflows through your own connected MCP-OpenStack-Ops server.

    Free

    Security scanned

    About this skill

    Give your AI agent structured OpenStack investigation and resource-operation workflows through your own connected MCP-OpenStack-Ops server.

    The skill combines 92 individual tool guides with six operational playbooks. Tool entries include full source-matched signatures, registration conditions, input requirements and practical interpretation guidance, organized into eight service groups:

    • Compute and server resources: inventory, state, events, attachments and lifecycle.
    • Networking: networks, subnets, routers, security groups and floating-IP bindings.
    • Block storage: volumes, snapshots and supported attachment/resource operations.
    • Images: image records, visibility, sharing and metadata.
    • Identity and project access: visible projects, users, roles and related resources.
    • Monitoring, quotas and services: resource evidence, limits and usage inspection.
    • Heat orchestration: stack review and authorized template-based operations.
    • Octavia load balancing: LBs, listeners, pools, members, monitors and L7 routing.

    The playbooks guide instance triage, connectivity investigation, storage maintenance preparation, load-balancer analysis, quota/usage review, and image/identity/Heat work. They distinguish cloud state from application health, API submission from completion, and a bounded result page from a full inventory.

    Inspection comes first. Changes require the environment to permit the operation and explicit user authorization for the concrete resources, action and impact. The agent establishes bulk targets before execution and preserves partial failures. Changes can disrupt workloads, alter access or public exposure, delete data and incur resource costs; the skill requires those effects to be covered by authorization and checks the result afterward.

    This product supplies instructions, not infrastructure or cloud privileges. Actual action support depends on the deployed MCP version, wrapper inputs, OpenStack services and policy. The agent does not bypass missing tools or enable writes automatically. The skill includes implementation-specific cautions and requires live-schema checks rather than promising every advertised action works.

    Mutation requires ALLOW_MODIFY_OPERATIONS=true. set_quota, set_server_dump, and snapshot restore are currently unimplemented or broken in the baseline. Administrative scope depends on API policy.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 19 days ago

    • Free with an Agensi account

    Requires

    OpenStack Ops MCP
    OpenStack
    OpenStack policy permissions
    Python 3.12+
    OpenStack Heat
    OpenStack Octavia
    Python

    Permissions

    Network Access
    Listed19 days ago

    What's inside

    Frequently Asked Questions

    No. Connect your own MCP-OpenStack-Ops server to your own cloud. This package contains a self-contained SKILL.md, not installation software or credentials.

    No. There are 41 unconditional queries and 51 conditional tools in the reviewed source. Conditional registration depends on server configuration; even read actions within those tools can be absent. A registered tool does not guarantee that every action in its documentation works in the deployed environment.

    It covers instance-query pagination differences, missing search forwarding, general-quota wrapper/helper mismatch, unavailable destination inputs for some restore paths, list-only volume-group/QoS helpers, unsupported server dumps, and differences between advertised and implemented flavor/L7/monitor actions. Use live schemas and actual results when deployed versions change. Service log show in the reviewed baseline returns descriptive guidance, not actual log lines.

    No. A diagnosis does not authorize mutation. The workflow requires approved concrete targets/action/impact and a server policy that permits it. Mutation wrappers may immediately act on discovered targets; the skill requires separate read-only target discovery first. Batches can partially succeed and are not atomic.

    No. Asynchronous resource operations require follow-up inspection. The skill separates submission, observed terminal state and verified postconditions. It does not claim guest application health, backup restoreability or complete inventory from a control-plane status or a limited response page.

    New in DevOps & Cloud

    Free