[](https://www.claudepluginhub.com/plugins/jpantsjoha-googlecloud-plugin?ref=badge) [](https://agent-plugins.org/specification) [](https://agentskills.io/specification) # GoogleCloud Plugin > **Install once, and your coding agent gains a full Google Cloud delivery team — a vendor-objective solution designer, a GCP architect, an agentic-systems architect, plus security, SRE, and QA — wired into a design-first, security-first delivery gate.** > **All the GCP skills. All the MCP servers available today. Assembled to fast-track the right architecture for your requirements.** Skill libraries for Google Cloud are everywhere. What they miss is the operating contract: who designs before anyone writes code, who signs off on least-privilege before anything deploys, what "done" means when an agent ships to a billable cloud. This plugin ships that contract, with the GCP know-how built in. One install gives an agent (Claude, Codex, or Antigravity) six personas, eleven service skills, and a research pipeline that keeps every reference current. A plugin that goes stale in silence is worse than none. **Compliant with [Agent Plugins 1.0.0](https://agent-plugins.org/specification)** — the open, vendor-neutral packaging standard co-maintained by Amazon, Cursor, Google, Microsoft, OpenAI and Vercel ([announcement](https://developers.googleblog.com/agent-plugins-package-your-skills-tools-and-more/)) — and with the [Agent Skills specification](https://agentskills.io/specification) it references. One directory layout works across conformant clients, rather than a bespoke wrapper per harness. Enforced in CI by `make spec` and verified against the official reference validator; see [Standards and compliance](#standards-and-compliance). The plugin is **GCP-only and opinionated by design**. It covers the services a builder reaches for — Firebase, Cloud Run, BigQuery, GKE — and the agent frameworks now built on top: ADK, and the MCP / A2A / AP2 protocol stack, deployed on the Gemini Enterprise Agent Platform. Refine it for your team. Keep the contract coherent as you do. > **Speed is easy. Safe speed is engineered.** --- ## Install One plugin, four coding assistants. Skills live in `skills/`; each harness reads its own manifest. **Claude Code** ```text /plugin marketplace add jpantsjoha/googlecloud-plugin /plugin install googlecloud-plugin@googlecloud-plugin-marketplace ``` **Antigravity (Gemini)** ```bash agy plugin install https://github.com/jpantsjoha/googlecloud-plugin ``` **Kimi Code** ```text /plugins install https://github.com/jpantsjoha/googlecloud-plugin ``` **Codex** — no install step. Codex discovers `.agents/skills/` natively and reads `AGENTS.md` as its always-on adapter. Clone the repo and Codex picks up the contract: ```bash git clone https://github.com/jpantsjoha/googlecloud-plugin ``` After cloning to work on the plugin itself, install the pre-commit gate once: ```bash make hooks # runs `make gate` (validate + manifest + mermaid + lint + test) on every commit ``` The plugin running in Antigravity — asked what it is and how it delivers value, the agent introduces the team and the delivery gate:

--- ## The Solution → Delivery Workflow Every piece of work flows through sequential gates. No gate opens until the one before it clears. Human intent enters at the top; a human authorises production at the bottom. The AI personas scale the execution in between — they never grant their own authority. ```mermaid %%{init: {'theme':'neutral', 'sequence': {'useMaxWidth': true, 'diagramMarginX': 24, 'diagramMarginY': 18, 'actorMargin': 44, 'boxMargin': 15, 'boxTextMargin': 8, 'noteMargin': 12, 'messageMargin': 42}}}%% sequenceDiagram autonumber actor Req as Requirement participant Sol as solution-designer participant Arch as gcp-architect +
agent-architect participant Sec as gcp-security participant Impl as Service Skills participant QA as gcp-qa actor Ops as gcp-ops rect rgb(238, 242, 255) note over Req,Sol: Solution Gate — is this even GCP? (skip if unambiguously GCP-only) Req->>Sol: 1. Submit intent / problem Sol-->>Arch: 2. Master HLD + vendor scope (hands GCP slice down) end rect rgb(237, 233, 254) note over Arch: GCP Design Gate — design before code Arch->>Arch: 3. GCP HLD/LLD + ADRs (agentic: topology, protocols, model, eval set) end rect rgb(252, 231, 243) note over Arch,Sec: Security Gate — reviews the design, not the diff Arch->>Sec: 4. Submit design Sec-->>Impl: 5. Least-privilege IAM, no secrets, WAF pillar — pass or BLOCK end rect rgb(240, 253, 244) note over Sec,Impl: Implementation — known-good GCP patterns, cost-warned Impl->>Impl: 6. Build on google-cloud patterns, warn before every billable action end rect rgb(254, 249, 231) note over Impl,QA: Quality Gate — receipts, not polish Impl->>QA: 7. Submit candidate QA-->>Ops: 8. make gate (all checks) + agent-eval — pass or BLOCK end rect rgb(224, 242, 254) note over QA,Ops: Operational Readiness Gate — no dark production Ops->>Ops: 9. SLO + alerting + runbook present? Ops-->>Req: 10. Authorise release & emit status end ``` ### Gate ownership | Gate | Owner | Blocks release when | | :--- | :--- | :--- | | **Solution** | `solution-designer` | Vendor choice unclear, or GCP scope not delimited (skips if GCP-only) | | **GCP Design** | `gcp-architect` (+ `agent-architect` if agentic) | No GCP HLD, or HLD contradicts the solution design | | **Security** | `gcp-security` | IAM over-permissive, secrets in code, GCP risks unaddressed | | **Quality** | `gcp-qa` (+ `agent-architect` agent-eval) | `make gate` fails, acceptance criteria unmet, agent eval fails | | **Operational Readiness** | `gcp-ops` | No SLO, alerting, or runbook for a production workload | The rule underneath every gate: **infer intent, never infer permission.** An AI persona drafts, designs, and validates; a named human authorises anything consequential, billable, or irreversible. --- ## What you install **Six personas** — the team: | Persona | Tier | Owns | | :--- | :--- | :--- | | `solution-designer` | 1 · vendor-agnostic | The overarching solution; objective vendor selection across GCP / AWS / Azure | | `gcp-architect` | 2 · GCP | GCP infrastructure design — HLD/LLD, ADRs, IAM and network topology | | `agent-architect` | 2 · GCP agentic | Agentic systems on GEAP — ADK, Agent Runtime, MCP/A2A/AP2, and agent evaluation | | `gcp-security` | 3 | Least-privilege IAM, secrets, Well-Architected security pillar | | `gcp-ops` | 3 | SLOs, alerting, runbooks, operational readiness | | `gcp-qa` | 3 | Validation, linting, smoke tests, the quality gate | **Eight first-wave service skills:** Cloud Run · GKE · IAM · BigQuery · Cloud Storage · Vertex AI · Networking · Logging/Monitoring. **Three cross-cutting skills:** `mcp-servers` (Google-managed + self-hosted MCP setup) · `well-architected` (the six-pillar framework) · `terraform-gcp` (CFT + provider patterns). --- ## Standards and compliance This plugin conforms to the open, vendor-neutral packaging standards — not to any single vendor's plugin format. | Standard | Version | What conforms | How it's verified | |---|---|---|---| | [Agent Plugins](https://agent-plugins.org/specification) | 1.0.0 | `plugin.json`, `mcp.json`, the `skills/` layout | `make spec` + [`plugin.schema.json`](https://agent-plugins.org/schemas/1.0.0/plugin.schema.json) / [`mcp.schema.json`](https://agent-plugins.org/schemas/1.0.0/mcp.schema.json) | | [Agent Skills](https://agentskills.io/specification) | current | all 17 `SKILL.md` files | `make validate` + the official [`skills-ref`](https://github.com/agentskills/agentskills/tree/main/skills-ref) reference validator | | [Model Context Protocol](https://modelcontextprotocol.io) | stdio transport | `mcp.json` server declarations | `make spec` transport + placeholder checks | Agent Plugins 1.0.0 is co-maintained by Amazon, Cursor, Google, Microsoft, OpenAI and Vercel. Conforming to it means one directory layout works across conformant clients instead of a bespoke wrapper per harness. The per-harness manifests (`.claude-plugin/`, `.kimi-plugin/`, `gemini-extension.json`, `.agents/`) remain alongside it, so existing installs are unaffected. Verify it yourself, without trusting this repo's own scripts: ```bash uvx --from "git+https://github.com/agentskills/agentskills.git#subdirectory=skills-ref" \ skills-ref validate skills/cloud-run uvx check-jsonschema --schemafile https://agent-plugins.org/schemas/1.0.0/plugin.schema.json plugin.json ``` See [ADR-007](architecture/decisions/ADR-007-agent-plugins-conformance.md) for the adoption decision and its trade-offs. --- ## Validation The plugin validates itself. Every skill conforms to a machine-readable contract; every reference URL is checked live; every source carries a retrieval date and content hash for audit. ```bash make gate # the full gate: spec + validate + manifest + mermaid + lint + test make spec # Agent Plugins 1.0.0 conformance (plugin.json + mcp.json + skills/) make validate # SKILL.md frontmatter vs the Agent Skills specification make manifest # per-harness manifests present and consistent make lint # every reference URL resolves (HTTP 200) make test # skill smoke tests make check # freshness: content-hash drift vs live GCP docs make crawl # refresh the research corpus (weekly) ``` Freshness is a feature, not a chore: `make check` runs weekly and flags any GCP doc that has drifted since it was last captured, so the plugin never quietly rots. --- ## About the author Created and maintained by **Jaroslav Pantsjoha (JP)** — Technical Director and Enterprise Agent Solution Architect at Cognizant, Google Developer Expert (Google Cloud), speaker, and author of *Building the Agentic Enterprise on Google Cloud* (Packt). I built this to make my own Google Cloud work repeatable, then to share it. The model is the easy part; the durable engineering is the harness around it — the skills, rules, MCP servers, gates, and evals. GCP patterns outlast the week's model release, so they are worth encoding once and reusing. This plugin is that baseline aimed squarely at Google Cloud — the GCP companion to my [`join-the-team`](https://github.com/jpantsjoha/ai-native-developer-experience) harness. Part of the **#HarnessEngineering** body of work — the engineering discipline behind the Agentic Enterprise. - LinkedIn: [uk.linkedin.com/in/johas](https://uk.linkedin.com/in/johas) - Google Developer Expert: [me.developers.google.com/u/jpantsjoha](https://me.developers.google.com/u/jpantsjoha) --- ## Licence Open source under the [MIT License](LICENSE) — free for personal and commercial use, modification, and redistribution. --- *This is an independent, community project. It is not affiliated with, sponsored by, or endorsed by Google LLC. "Google Cloud", "GCP", "Gemini", "Vertex AI", and related names are trademarks of Google LLC, used here only to describe what the plugin works with.*