home / skills / athola / claude-night-market / gemini-delegation

gemini-delegation skill

safe

/plugins/conjure/skills/gemini-delegation

This skill manages Gemini CLI delegation through delegation-core, enabling authentication, quota checks, and batch analysis using Gemini's large context window.

npx playbooks add skill athola/claude-night-market --skill gemini-delegation

Review the files below or copy the command above to add this skill to your agents.

Files (2)

SKILL.md

3.8 KB

---
name: gemini-delegation
description: Gemini CLI delegation workflow implementing delegation-core for Google's
  Gemini models. Use when delegation-core selected Gemini, need Gemini's 1M+ token
  context window, batch processing required. Do not use when deciding which model
  to use (use delegation-core first), gemini CLI not installed.
category: delegation-implementation
tags:
- gemini
- cli
- delegation
- google
- large-context
dependencies:
- delegation-core
tools:
- gemini-cli
usage_patterns:
- gemini-cli-integration
- large-context-analysis
- batch-processing
complexity: intermediate
estimated_tokens: 600
progressive_loading: true
modules:
- modules/gemini-specifics.md
references:
- delegation-core/../../leyline/skills/authentication-patterns/SKILL.md
- delegation-core/../../leyline/skills/quota-management/SKILL.md
- delegation-core/../../leyline/skills/usage-logging/SKILL.md
- delegation-core/../../leyline/skills/error-patterns/SKILL.md
---
## Table of Contents

- [Overview](#overview)
- [When to Use](#when-to-use)
- [Prerequisites](#prerequisites)
- [Delegation Flow](#delegation-flow)
- [Quick Start](#quick-start)
- [Basic Command](#basic-command)
- [Save Output](#save-output)
- [Shared Patterns](#shared-patterns)
- [Gemini-Specific Details](#gemini-specific-details)
- [Exit Criteria](#exit-criteria)


# Gemini CLI Delegation

## Overview

This skill implements `conjure:delegation-core` for the Gemini CLI. It provides Gemini-specific authentication, quota management, and command construction using shared patterns.

## When To Use
- After `Skill(conjure:delegation-core)` determines Gemini is suitable
- When you need Gemini's large context window (1M+ tokens)
- For batch processing, summarization, or pattern extraction tasks
- If the `gemini` CLI is installed and authenticated

## When NOT To Use

- Deciding which model
  to use (use delegation-core first)
- Gemini CLI not installed
- Deciding which model
  to use (use delegation-core first)
- Gemini CLI not installed

## Prerequisites

**Installation:**
```bash
# Verify installation
gemini --version

# Check authentication
gemini auth status

# Login if needed
gemini auth login

# Or set API key
export GEMINI_API_KEY="your-key"
```
**Verification:** Run the command with `--help` flag to verify availability.

## Delegation Flow

Implements standard delegation-core flow with Gemini specifics:

1. `gemini-delegation:auth-verified` - Verify Gemini authentication
2. `gemini-delegation:quota-checked` - Check Gemini API quota
3. `gemini-delegation:command-executed` - Execute via Gemini CLI
4. `gemini-delegation:usage-logged` - Log Gemini API usage

## Quick Start

### Basic Command
```bash
# File analysis
gemini -p "@path/to/file Analyze this code"

# Multiple files
gemini -p "@src/**/*.py Summarize these files"

# With specific model
gemini --model gemini-2.5-pro-exp -p "..."

# JSON output
gemini --output-format json -p "..."
```
**Verification:** Run the command with `--help` flag to verify availability.

### Save Output
```bash
gemini -p "..." > delegations/gemini/$(date +%Y%m%d_%H%M%S).md
```
**Verification:** Run the command with `--help` flag to verify availability.

## Shared Patterns

This skill uses shared modules from delegation-core:
- **Authentication**: See `delegation-core/../../leyline/skills/authentication-patterns/SKILL.md`
- **Quota Management**: See `delegation-core/../../leyline/skills/quota-management/SKILL.md`
- **Usage Logging**: See `delegation-core/../../leyline/skills/usage-logging/SKILL.md`
- **Error Handling**: See `delegation-core/../../leyline/skills/error-patterns/SKILL.md`

## Gemini-Specific Details

For Gemini-specific models, CLI options, cost reference, and troubleshooting, see `modules/gemini-specifics.md`.

## Exit Criteria
- Authentication confirmed working
- Quota checked and sufficient
- Command executed successfully
- Usage logged for tracking

Overview

This skill implements the conjure:delegation-core workflow for the Gemini CLI, adding Gemini-specific authentication, quota checks, and command construction. It enables use of Gemini models with very large context windows and supports batch processing and structured output. Use this skill only after delegation-core chooses Gemini and when the gemini CLI is installed and authenticated.

How this skill works

The skill verifies local Gemini authentication and checks API quota before constructing and running gemini CLI commands tailored to large-context or batch workloads. It executes commands, captures JSON or plain output to files, and logs usage for tracking and billing. Built-in patterns handle errors and ensure the delegation flow reports auth, quota, execution, and usage logging events.

When to use it

After delegation-core determines Gemini is the selected provider
When you need Gemini’s 1M+ token context window for long documents or memory-heavy tasks
For batch processing, bulk summarization, or pattern extraction across many files
When the gemini CLI is installed and authenticated on the host
When you need structured JSON output or to persist command results for auditing

Best practices

Run gemini --version and gemini auth status before delegating to confirm installation and login
Use --output-format json for machine-friendly results and direct output to timestamped files for traceability
Check quota proactively to avoid mid-job failures and log usage after each delegation
Keep commands focused per run (single analysis or a defined batch) to make cost and performance predictable
Prefer delegation-core to decide model selection; use this skill only after Gemini is chosen

Example use cases

Summarize a codebase using gemini -p "@src/**/*.py Summarize these files" and save JSON output for downstream tooling
Analyze a very large document that exceeds other models’ context by using a Gemini model with 1M+ token support
Run batch pattern extraction across many source files and export results to delegations/gemini/<timestamp>.md for review
Execute model-specific experiments (specific gemini version or pro experimental flags) via --model and log usage for quota accounting

FAQ

What prerequisites are required?

Install the gemini CLI, verify it with gemini --version, and authenticate via gemini auth login or set GEMINI_API_KEY.

How do I save command output?

Redirect output to a file, e.g. gemini -p "..." > delegations/gemini/$(date +%Y%m%d_%H%M%S).md or use --output-format json.