jeremylongshore/tons-of-skills-marketplace / plugins/ai-ml/jeremy-gcp-starter-examples/skills/gcp-examples-expert
gcp-examples-expert
VerifiedGenerate production-ready Google Cloud code examples from official repositories including ADK samples, Genkit templates, Vertex AI notebooks, and Gemini patterns. Use when asked to "show ADK example" or "provide GCP starter kit". Trigger with relevant phrases based on skill purpose.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 6.1 KB |
| references/ARD.md | markdown | 5.2 KB |
| references/PRD.md | markdown | 4.2 KB |
| references/best-practices-applied.md | markdown | 1.2 KB |
| references/code-example-categories.md | markdown | 4.1 KB |
| references/errors.md | markdown | 4.4 KB |
| references/example-interactions.md | markdown | 1.5 KB |
| references/examples.md | markdown | 12.9 KB |
| references/workflow.md | markdown | 972 B |
| scripts/create-example.shexecutable | script | 2.0 KB |
references/ARD.md
View on GitHub# ARD: GCP Examples Expert
> Part of [Tons of Skills](https://tonsofskills.com) by [Intent Solutions](https://intentsolutions.io) | [jeremylongshore.com](https://jeremylongshore.com)
## System Context
The GCP Examples Expert maps developer requirements to the correct Google Cloud AI repository and generates production-ready code adapted from official patterns. It operates as a code generation skill that bridges six official GCP AI repositories.
```
Developer Request ("Build a RAG system on GCP")
↓
[GCP Examples Expert]
├── Maps to: Framework selection (Genkit, ADK, Vertex AI, etc.)
├── Sources: 6 official GCP repositories
├── Adapts: language, security, monitoring, deployment
└── Generates: runnable code + deployment config + IaC
↓
Production-Ready Code Package
├── Application code (adapted from source repo)
├── Deployment config (Cloud Run / Functions / Agent Engine)
├── Terraform module (infrastructure)
└── Documentation (source citations, cost estimates)
```
## Data Flow
1. **Input**: User request describing the AI use case, preferred language, and deployment target. May include constraints like model preference, cost budget, or compliance requirements.
2. **Processing**: Classify the request into one of six framework categories. Select the matching code pattern from the categorized reference. Adapt to the target language. Layer on security (IAM, VPC-SC, Secret Manager), monitoring (OpenTelemetry, dashboards), and deployment configuration. Generate Terraform when infrastructure provisioning is needed.
3. **Output**: Complete code package with application files, deployment config (YAML/JSON), Terraform module, environment variable template, monitoring setup, cost estimate, and source repository citations.
## Key Design Decisions
| Decision | Choice | Rationale |
|----------|--------|-----------|
| Six-category framework taxonomy | ADK, Starter Pack, Genkit, Vertex AI, Gen AI, AgentSmithy | Covers the complete GCP AI landscape; each category maps to a distinct official repo |
| Security-by-default | Every example includes IAM, Secret Manager, and VPC-SC | Production examples must not teach insecure patterns; developers copy-paste examples |
| Source attribution | Cite the specific repo and pattern for every adaptation | Enables developers to find updates and deeper documentation; maintains trust |
| Model selection guidance | Gemini 2.5 Flash for throughput, Pro for reasoning | Prevents cost surprises; Flash handles 80% of use cases at 10x lower cost |
| Multi-target deployment | Generate configs for Cloud Run, Firebase Functions, and Agent Engine | Different deployment targets suit different flow types; developers choose |
| Language adaptation | Transform patterns to TypeScript, Python, or Go | Each repo has language-specific conventions; adapt rather than translate literally |
| IaC inclusion | Terraform templates alongside application code | Production deployments need reproducible infrastructure, not just application code |
## Tool Usage Pattern
| Tool | Purpose |
|------|---------|
| Read | Inspect existing project files, requirements, and configuration to adapt examples to the user's codebase |
| Write | Create new application files, deployment configs, Terraform modules, and environment templates |
| Edit | Patch existing code to integrate GCP patterns, update dependencies, or add monitoring |
| Grep | Search for existing framework usage, import patterns, and configuration in the target project |
| Glob | Discover project structure to determine where generated code should be placed |
| Bash(cmd:*) | Run dependency installation, gcloud commands, firebase CLI, and validation scripts |
## Error Handling Strategy
| Error Class | Detection | Recovery |
|------------|-----------|----------|
| Framework mismatch | User request doesn't clearly map to a single category | Ask clarifying questions; present the two closest matches with trade-offs |
| API not enabled | `gcloud services list` missing required APIs | Provide the exact `gcloud services enable` command for each missing API |
| Region unavailability | Requested model not available in specified region | Suggest `us-central1` or `europe-west4` where Gemini models are available |
| Dependency conflict | Incompatible SDK versions in existing project | Pin versions explicitly; provide a clean `requirements.txt` or `package.json` |
| Quota exceeded | Rate limit on Vertex AI prediction endpoints | Include exponential backoff in generated code; provide quota increase request link |
## Extension Points
- Additional framework categories: add new GCP AI repos (e.g., MLOps pipelines) as they become official
- Language expansion: extend beyond TypeScript/Python/Go when Genkit or ADK adds language support
- Custom security profiles: allow enterprise users to specify compliance frameworks (SOC2, HIPAA) for stricter defaults
- Cost calculator integration: generate detailed cost projections based on expected query volume and model selection
- Template versioning: pin examples to specific SDK versions with automated upgrade paths
- Interactive selection: when multiple frameworks match, present a comparison table for the user to choose
- Batch generation: create examples for multiple frameworks from a single requirements document
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| writing-skillsobra/superpowersUse when creating new skills, editing existing skills, or verifying skills work before deployment | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| xlsxanthropics/skillsUse this skill any time a spreadsheet file is the primary input or output. This means any task where the user wants to: open, read, edit, or fix an existing .xlsx, .xlsm, .xltx, .csv, or .tsv file (e.g., adding columns, computing formulas, formatting, charting, cleaning messy data); create a new spreadsheet from scratch or from other data sources; or convert between tabular file formats. Trigger especially when the user references a spreadsheet file by name or path — even casually (like "the xlsx in my downloads") — and wants something done to it or produced from it. Also trigger for cleaning or restructuring messy tabular data files (malformed rows, misplaced headers, junk data) into proper spreadsheets. The deliverable must be a spreadsheet file. Do NOT trigger when the primary deliverable is a Word document, HTML report, standalone Python script, database pipeline, or Google Sheets API integration, even if tabular data is involved. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| shipping-and-launchaddyosmani/agent-skillsPrepares production launches. Use when preparing to deploy to production, or when asking what needs to be in place before shipping. Use when you need a pre-launch checklist, when setting up monitoring, when planning a staged rollout, or when you need a rollback strategy. | Stars: 119k | Updated | Claude CodeCodex | Verified | Install |
| childhoodthedotmack/claude-memBabysit a shipped release after merge, publication, or deployment. Verify published packages, production behavior, installed plugins, and late CI or user reports until the observation window is clean; use babysit for PR readiness before release. | Stars: 118k | Updated | Claude CodeCodex | Verified | Install |