What is litellm-vertex-gemini-local-gateway?

A guide for building a local LiteLLM gateway that connects Google Vertex AI Gemini to tools such as Claude Code and OpenClaw. A gateway is a local service that receives requests and forwards them to an AI model.

Setting up the gateway, connecting it to Vertex AI, configuring Claude Code or OpenClaw, starting it automatically, and verifying its API endpoints.

It provides one local endpoint for Gemini and can keep that service running after macOS login without disturbing existing configurations.

litellm-vertex-gemini-local-gateway — Skill for Claude Code and Codex

Install

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.

Any agent

npx skills add davidtoby/agent-skills --skill litellm-vertex-gemini-local-gateway

Clone the repo

git clone --depth 1 https://github.com/davidtoby/agent-skills

Made for: Claude Code, Codex.

Wrote 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.

README.md

[![agentmods](https://agentmods.dev/badge/skills/davidtoby/agent-skills/litellm-vertex-gemini-local-gateway/github.svg)](https://agentmods.dev/skills/davidtoby/agent-skills/litellm-vertex-gemini-local-gateway)

Your own site

<a href="https://agentmods.dev/skills/davidtoby/agent-skills/litellm-vertex-gemini-local-gateway"><img src="https://agentmods.dev/badge/skills/davidtoby/agent-skills/litellm-vertex-gemini-local-gateway/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

Your own site · 80×15

<a href="https://agentmods.dev/skills/davidtoby/agent-skills/litellm-vertex-gemini-local-gateway"><img src="https://agentmods.dev/badge/skills/davidtoby/agent-skills/litellm-vertex-gemini-local-gateway.svg" alt="Reviewed on agentmods" width="80" height="20"></a>

Per session 105 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.

When invoked 2,448 The whole file, excluding the scripts and references it only reads on demand.

Security scan D 3 findings. A grade says what 26 rules found in the file — not that it is safe.

Origin original No closer match found in the catalogue.

Token cost

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 5.1	$0.00105	$0.02448
Opus 5	$0.00053	$0.01224
Sonnet 5	$0.00021	$0.00490
Haiku 4.5	$0.00011	$0.00245

Measured 8d ago against content hash 57aec89c2762, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-11, from the pricing page.

Security

Grade D, and why

litellm-vertex-gemini-local-gateway scanned grade D with 3 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 8d ago.

The scan reads SKILL.md. This mod also ships 5 executable files (scripts/claude-direct-wrapper.sh, scripts/claude-gemini-wrapper.sh, scripts/claude-sonnet-direct-wrapper.sh, …), listed below but not scanned — reading those needs a real analyzer, not pattern matching.

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.

Reaches for credential fileshighPrivilege escalation

SSH keys, cloud credentials, git-credentials, .npmrc, /etc/shadow: reading these is how a config file becomes a credential leak.

- `~/.config/gcloud/application_default_credentials.json`

Reads agent configuration directoriesmediumAgent snooping

.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.

Back up `~/.claude/settings.json`, then set:

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

curl -sS -H "Authorization: Bearer $LITELLM_MASTER_KEY" http://127.0.0.1:4000/v1/models

skills/litellm-vertex-gemini-local-gateway/SKILL.md · 270 lines

How it starts

The opening of the file, as written. The whole thing — 270 lines — stays where its author put it; the contents beside it link to each section on GitHub.

LiteLLM Vertex Gemini Local Gateway

Build a local LiteLLM deployment on macOS that:

uses Vertex AI ADC for Gemini
exposes an Anthropic-compatible endpoint such as http://127.0.0.1:4000
auto-starts after login via LaunchAgent and stays resident in the background
lets Claude Code use Gemini through LiteLLM
lets OpenClaw use the same local gateway as a model option without disturbing the existing default unless requested

Quick start

Confirm prerequisites:
- macOS
- uv
- python3
- Vertex AI ADC already available or obtainable
- claude installed if Claude Code integration is requested
- openclaw installed if OpenClaw integration is requested
Read references/fresh-macos-runbook.md first.
Install LiteLLM with Google support:
- uv tool install 'litellm[proxy,google]'
Generate or hand-write the gateway project files using scripts/render_gateway_bundle.py or references/config-templates.md.
Put secrets only in .env and never echo them back to the user.
Install/load the LaunchAgent and verify port listening.
Verify both /v1/models and /v1/messages.
Wire Claude Code.
Wire OpenClaw.
Report changed files, test commands, rollback paths, and any provider-specific caveats.

What makes this setup fragile

The most important pitfalls are:

LiteLLM can appear healthy on /v1/models but still fail real inference until google extras are installed.
Claude Code must point ANTHROPIC_BASE_URL at the LiteLLM root, not /v1.
Python/system proxy settings can break loopback traffic unless NO_PROXY=127.0.0.1,localhost,::1 is forced.
OpenClaw per-run gateway --model overrides may be unauthorized in some local setups; in that case use --local with the explicit provider/model id or switch the default with openclaw models set.

Read references/troubleshooting.md before improvising.

Workflow

1. Confirm prerequisites and choose paths

Choose or confirm:

PROXY_DIR — local project directory, e.g. ~/GitHub-Codebase/litellm-vertex-proxy
LITELLM_HOST — usually 127.0.0.1
LITELLM_PORT — usually 4000
LaunchAgent label — e.g. com.example.litellm-vertex-proxy
Vertex project id
Vertex location, usually global
LiteLLM model alias, e.g. gemini-3.1-pro-preview

Read the full file on GitHub · 270 lines

Files

What ships with it

12 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.

Changes

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.

8d ago First seen · 270 lines · 105 tokens per session scan D 57aec89c2762

Subscribe to this mod's changes

litellm-vertex-gemini-local-gateway is a skill published in the GitHub repository davidtoby/agent-skills (10 stars, last pushed 1mo ago), licensed MIT. It adds 105 tokens to every session and 2,448 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it D with 3 findings (reaches for credential files, reads agent configuration directories, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.

Cloud Data and AI DevOps and cloud agent ai-agents automation claude openclaw skills

azure-data-factory

Expert knowledge for Azure Data Factory development including troubleshooting, best practices, decision making, architecture & design patterns, limits & quotas, security, configuration, integrations & coding patterns, and deployment. Use when building ADF pipelines, mapping data flows, SSIS IR/SHIR, CI/CD deployments…

MicrosoftDocs/Agent-Skills · 127 tokens

azure-databricks

Expert knowledge for Azure Databricks development including troubleshooting, best practices, decision making, architecture & design patterns, limits & quotas, security, configuration, integrations & coding patterns, and deployment. Use when using Unity Catalog, Lakeflow pipelines, Genie/AI Runtime, Delta…

MicrosoftDocs/Agent-Skills · 124 tokens

azure-hdinsight

Expert knowledge for Azure HDInsight development including troubleshooting, best practices, decision making, architecture & design patterns, limits & quotas, security, configuration, integrations & coding patterns, and deployment. Use when working with HDInsight Spark/Hive/Kafka/HBase clusters, Ambari/Oozie pipelines…

MicrosoftDocs/Agent-Skills · 119 tokens

azure-microsoft-discovery

Expert knowledge for Azure Microsoft Discovery development including troubleshooting, best practices, decision making, architecture & design patterns, limits & quotas, security, configuration, integrations & coding patterns, and deployment. Use when building Discovery Engine shared sessions, ACR tool images, REST…

MicrosoftDocs/Agent-Skills · 78 tokens

azure-data-science-vm

Expert knowledge for Azure Data Science Virtual Machines development including troubleshooting, decision making, architecture & design patterns, security, configuration, integrations & coding patterns, and deployment. Use when managing DSVM images/tools, IaC deployment (Bicep/ARM), Key Vault secrets, MLflow, or…

MicrosoftDocs/Agent-Skills · 126 tokens

ehr-analysis

End-to-end EHR predictive modeling pipeline with PyHealth, covering dataset loading, task definition, model training, evaluation, calibration, and clinical interpretation.

zongtingwei/Bioclaw_Skills_Hub · 33 tokens

litellm-vertex-gemini-local-gateway