Developer Tools●Mid
A CLI tool for agentic code review and auto-fixing
Automates Codex-style review-fix loops, but multi-pass LLM review is table stakes now.
Big Brain
kenryu
303mo ago
Back to browse
GitHub Repository
Orchestrating coding agents for code review, verification and fixing via the ralph loop.
13 starsTypeScript
Ralph Review – OSS code review that loops fixes until no issues remain
by kenryu·Mar 5, 2026·1 point·0 comments
AI Analysis
●MidBig BrainShip It
Automated code review loop via agent ping-pong, but Cursor already does multi-turn fixing in context.
Strengths
- •Ralph Wiggum loop design (reviewer → fixer → verify → repeat) is genuinely clever; fresh context avoids first-agent bias
- •Flexible agent support (Codex, Claude, others) means not locked into one model; MIT license, open repo
- •Solves real developer friction: manual copy-paste between review and fix sessions felt tedious enough to justify automation
Weaknesses
- •Cursor, Continue, and GitHub Copilot all ship integrated fixing—agents don't need a CLI wrapper anymore
- •No benchmarks or examples of actual fixes; unclear if loop convergence is faster than human review + single fix pass
Category
Target Audience
Engineering teams wanting automated code review loops, developers using Cursor/Codex-like agents
Similar To
Cursor · GitHub Copilot · Continue IDE
Similar Projects
Security●●Solid
Nelson, a Ralph-like loop for finding vulnerabilities
Review mode filters false positives before escalating to humans.
Big BrainNiche Gem
SwellJoe
412mo ago
Developer Tools●●Solid
Automate Claude in a work->review loop with cook
Work-review-gate loop automates Claude iteration, but templated agentic loops already exist at scale.
Ship ItSolve My Problem
staticvar
203mo ago
Developer Tools●●●Banger
Triplecheck – Review your code free with local LLMs
Loop-based multi-pass review with auto-patching beats single-pass LLM lingers; local compute reframes economics.
WizardryZero to OneNiche Gem
raullen
103mo ago
AI/ML●●●Banger
CRTX – AI code gen that tests and fixes its own output (OSS)
Ditched multi-model bloat, proved single model + local test loop beats expensive debate.
Big BrainShip ItZero to One
johnnycash926
213mo ago
AI/ML●●Solid
Brute-force startup ideation with the Ralph Loop
The core trick is simple and effective: let an agent iterate questions against a defined domain overnight and surface hundreds of candidly-annotated ideas you can scan through later. It nails the “fire-and-forget” idea dump and domain steering (tell it to focus on agencies or cybersecurity and it pivots), but it’s still essentially a convenience wrapper around an existing agent pattern — useful for volume and pattern recognition, less convincing on long-term validation or downstream filtering.
Rabbit HoleNiche Gem
bothlabs
203mo ago