Agent Development Kit for Go is a Go framework for building, evaluating, deploying, and coordinating AI agents through code. Developers use it to define agent logic, connect tools, compose multi-agent systems, and deploy applications in cloud environments, while retaining compatibility with models and frameworks beyond Gemini. The catalogue entries provide skills and instructions for working with ADK Go.
Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/google/adk-go/adk-sample-creatornpx skills add google/adk-go --skill adk-sample-creatorgit clone --depth 1 https://github.com/google/adk-goWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/google/adk-go/adk-sample-creator)<a href="https://agentmods.dev/skills/google/adk-go/adk-sample-creator"><img src="https://agentmods.dev/badge/skills/google/adk-go/adk-sample-creator.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00094 | $0.01584 |
| Opus 5 | $0.00047 | $0.00792 |
| Sonnet 5 | $0.00019 | $0.00317 |
| Haiku 4.5 | $0.00009 | $0.00158 |
Grade A, and why
adk-sample-creator scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 140 lines — stays where its author put it; the contents beside it link to each section on GitHub.
ADK Go sample creator
An example earns its place by teaching one thing faster than the package docs can. It must run. Two families set the standard, and which one you are writing decides several of the choices below:
- Engine samples —
examples/workflow/*. They exercise ADK itself, need nothing beyond the binary and sometimes a model, and their subject is the shape of a graph. - Integration samples —
examples/agentregistry/*. They talk to a live external service, so they carry cloud prerequisites, credentials, failures worth unwrapping, and output that depends on the reader's own project.
Copy the shape of the nearer family rather than inventing one.
examples/README.md documents an older launcher API (ParseAndRun,
FormatSyntax). Follow the templates here, not that file.
Layout
examples/<area>/<name>/
├── main.go
└── README.md
lower_snake directory names (hitl_simple), nested by theme once a family
grows (routing/llm, dynamic/hitl). When the area keeps an index table —
examples/workflow/README.md and examples/agentregistry/README.md do — add a
row with the same columns as its neighbours, including LLM?. Most areas keep
none, and none is needed.
main.go
Copy templates/main.go.tmpl: it carries the Apache
2.0 header the goheader linter demands, the package doc line, and the launcher
wiring. context.Background() is correct there — main is an entrypoint, and
it stays banned in library code.
full.NewLauncher() serves console, which is the default, plus web and its
api, a2a, webui, pubsub and eventarc sublaunchers. When anything is
deferred, move the body into run(ctx) error and let main do the log.Fatal.
Each of these has cost a review round:
- Pick a versioned model, not a
-latestalias, in any sample whose README documents the Vertex AI path: Vertex does not serve the aliases, so the documented instructions would fail at model construction.gemini-3.5-flashanswers on both backends. - A required variable is read plainly and checked for empty, failing with a
message that names it;
cmp.Or(os.Getenv("X"), "default")is for the optional ones, and hides a missing value if you reach for it first. Watch for a variable a neighbouring library also reads:GOOGLE_CLOUD_LOCATIONsteers genai and the Agent Registry client, so a region meant for the model silently moves the catalog lookup. log.SetFlags(0)when the output is a transcript. Otherwise every line carries a timestamp, and a README that omits it is showing output the code cannot produce.log.Fatalonly frommain. Inside a body that defers a close, it skips the defer.- Render API failures as something actionable — unwrap the typed error and
keep the part that identifies the fix (see
explainin theagentregistrysamples), rather than printing the service's whole JSON envelope. - Comments say WHY. Never narrate the next line.
What ships with it
2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago First seen · 140 lines · 94 tokens per session scan A 8ed13d015490
adk-sample-creator is a skill published in the GitHub repository google/adk-go (8,745 stars, last pushed yesterday), licensed Apache-2.0. It adds 94 tokens to every session and 1,584 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-02.
Other skills, from other repositories
aegisops-ai-v2
/aegisops-ai — Autonomous Governance Orchestrator workflow skill. Use this skill when the user needs Autonomous DevSecOps & FinOps Guardrails. Orchestrates Gemini 3 Flash to audit Linux Kernel patches, Terraform cost drifts, and K8s compliance and the operator should preserve the upstream workflow, copied support…
go-expert
Expert-level Go development with Go 1.22+ features, concurrency, standard library, and production-grade best practices.
aegisops-ai
/aegisops-ai — Autonomous Governance Orchestrator workflow skill. Use this skill when the user needs Autonomous DevSecOps & FinOps Guardrails. Orchestrates Gemini 3 Flash to audit Linux Kernel patches, Terraform cost drifts, and K8s compliance and the operator should preserve the upstream workflow, copied support…
echo
Echo high-performance Go web framework. Use for Go APIs.
gin
Gin Go web framework with high performance and middleware. Use for Go APIs.
go-gin-generator
Generate production-ready Go projects using the Gin framework with modern best practices, multiple templates, database integration, authentication, and comprehensive features for building scalable web applications and microservices.