# Local Review

> Gemini-native local code review loop — iterates through review, fix, and commit cycles with 2M+ token grounding Use when this capability is needed.

- Skill: `tomevault-io/local-review` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add tomevault-io/local-review`
- Raw SKILL.md: https://api.skillmd.com/api/skills/tomevault-io/local-review/raw
- Safety review: pending (external: skill-scanner PASS, skillspector PASS)
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- Author: tomevault-io (https://skillmd.com/u/tomevault-io)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/tomevault-io/local-review

---


# Local Review Skill (Gemini Edition)

Iterative local code review loop that leverages Gemini's long context window to perform "whole-project" reviews of uncommitted changes, identifying structural flaws and architectural drift.

## Workflow
1. **Pre-Existing Scan**: Gemini runs a baseline check for lint/type errors, committing fixes separately to avoid mixing with feature work.
2. **Diff Review**: Gemini analyzes local changes against the base branch, classifying issues by keyword triggers (CRITICAL, MAJOR, MINOR, NIT).
3. **Fix Phase**: Gemini dispatches surgical fixes for identified issues in a priority-first loop.
4. **Iterative Convergence**: Continues reviewing and fixing until the codebase is clean or max iterations are reached.
5. **Commit Phase**: Automatically stages and commits fixes with standardized commit messages.

## Gemini Advantages
- **Multi-File Contextual Review**: Gemini sees all changed files in the context of the *entire* codebase, identifying breaking changes that local diff-based reviews miss.
- **Superior Intent Matching**: More accurately determines the severity of an issue by understanding its technical impact on the broader system architecture.
- **Intelligent Fingerprinting**: Better at deduplicating pre-existing issues against Linear tickets by analyzing the semantic meaning of the failure.

## Arguments
- `--uncommitted`: Only review uncommitted changes.
- `--since`: Base ref for diff (branch/commit).
- `--max-iterations`: Maximum review-fix cycles.
- `--required-clean-runs`: Consecutive clean runs required before passing.

---
> Source: [OmniNode-ai/omnigemini](https://github.com/OmniNode-ai/omnigemini) — distributed by [TomeVault](https://tomevault.io).
<!-- tomevault:4.0:skill_md:2026-05-23 -->

