Skill

gemini-large-analysis

This skill should be used when the user asks to analyze a large codebase, audit a full repository, review architecture across many files, map dependencies, or plan a migration involving more code than fits in Claude's context. Relevant when the user says "analyze this entire repo," "audit the architecture," or "this codebase is too big to review."

Install

npx claudepluginhub naluforge/geminicli-cc-plugin --plugin gemini

Tool Access

This skill uses the workspace's default tool permissions.

Preview

**Invoke using `/gemini-analyze` or `mcp__gemini__gemini_execute` with `include_directories` set to target paths.**

SKILL.md

Similar Skills

cache-components

Guides Next.js Cache Components and Partial Prerendering (PPR) with cacheComponents enabled. Implements 'use cache', cacheLife(), cacheTag(), revalidateTag(), static/dynamic optimization, and cache debugging.

cache-components

139.2k

claude-opus-4-5-migration

2 files

Migrates code, prompts, and API calls from Claude Sonnet 4.0/4.5 or Opus 4.1 to Opus 4.5, updating model strings on Anthropic, AWS, GCP, Azure platforms.

claude-opus-4-5-migration

83.2k

smart-explore

Performs token-optimized structural code search using tree-sitter AST parsing to discover symbols, outline files, and unfold code without reading full files.

claude-mem

65.8k

Stats

Stars1

Forks0

Last CommitApr 2, 2026

Actions

View Source View Plugin View on GitHub View README

Large Codebase Analysis with Gemini

Invoke using /gemini-analyze or mcp__gemini__gemini_execute with include_directories set to target paths.

When to Use Gemini vs Claude

Use Gemini when the analysis scope exceeds ~200K tokens — roughly a full repository or multiple large modules. Claude can handle focused analysis of individual files or small groups, but whole-repo audits, dependency trees, and cross-cutting architectural questions benefit from Gemini's 1M-token context window.

Recommended Approach

1. Scope First, Then Send

Before calling gemini_execute, understand what you're sending:

Use Glob to map the file tree: **/*.ts, **/*.py, etc.
Use Grep to find specific patterns: entry points, exports, imports
Estimate the token count (rough: 1 line ≈ 10 tokens)
If scope exceeds even Gemini's context, plan multiple focused passes

2. Parameter Selection

model: Always pro — codebase analysis requires deep reasoning
timeout_ms: 2400000 (40 minutes) — full-repo analysis can take 20-40 minutes
include_directories: Target specific directories rather than the entire repo when possible. This reduces noise and improves analysis quality.
max_turns: Increase if you want Gemini to use tools (file reading, etc.)

3. Breaking Into Passes

For truly massive codebases, break into focused passes:

Architecture pass: "Map the high-level module structure and dependencies"
Pattern pass: "Identify design patterns and anti-patterns in src/"
Quality pass: "Find code quality issues, dead code, and tech debt in lib/"
Synthesis: Combine findings from each pass

4. Cross-Reference with Local Inspection

Gemini's findings are a starting point. For each finding:

Verify with local file reads (the code may have changed)
Check git blame for context on why code is the way it is
Run relevant tests to confirm suspected issues

Output Format

Structure analysis output as:

Key findings — the most important observations
Architecture insights — module structure, dependency flow, coupling
Issues — with file:line references and severity
Recommendations — concrete, actionable next steps