
Agent Review Pipeline
- Updated April 17, 2026
- onchainyaotoshi/agent-review-pipeline
agent-review-pipeline is a plugin marketplace that adds an autonomous dual-engine code review pipeline with confidence-weighted consensus and optional bounded auto-fix.
About
agent-review-pipeline is a Claude Code plugin marketplace entry that ships one plugin: an autonomous code review pipeline for developers and small teams who want stronger pre-merge signal without manually juggling multiple review tools. It dispatches work asymmetrically across engines—Codex handles correctness and adversarial framings while Gemini runs a compound-engineering review path—so you get three parallel perspectives per iteration by default. Findings are fingerprinted, low-confidence noise is filtered, and agreement across sources boosts confidence. An optional bounded auto-fix loop can apply fixes across one to ten iterations, with a kill switch if the same issues repeat. Dry-run mode and defaults that keep commits and PR comments off make it safer to adopt on real repos. Install when you are in the ship phase and want repeatable, confidence-scored review coverage instead of a single generic linter pass.
- Asymmetric dual-engine dispatch: Codex runs ARP correctness plus adversarial framings (2 dispatches); Gemini runs its /c
- Three parallel review perspectives per iteration by default with fingerprinted findings and +0.15 confidence boost when
- Findings below 0.60 confidence dropped; bounded auto-fix loop (1–10 iterations) with loop-thrash kill switch
- --dry-run flag and safe-off defaults so commit and PR comment actions stay opt-in
- Single-plugin marketplace (agent-review-pipeline v6.5.0) focused on review automation, not general codegen
Agent Review Pipeline by the numbers
- Data as of Jul 7, 2026 (Skillselion catalog sync)
/plugin marketplace add onchainyaotoshi/agent-review-pipelineAdd your badge
Show developers this marketplace is listed on Skillselion. Paste this into your README.
| Last updated | April 17, 2026 |
|---|---|
| Repository | onchainyaotoshi/agent-review-pipeline ↗ |
What it does
Run autonomous multi-engine code reviews with confidence-weighted consensus and optional bounded auto-fix before you merge or ship.
Who is it for?
Best when you're shipping from Claude Code and want Codex-plus-Gemini review coverage on PRs before merge, with explicit controls on auto-fix and PR noise.
Skip if: Skip if you only need a fast linter in CI with no multi-model review budget, or projects that forbid external review dispatch entirely.
What you get
You get three structured review perspectives per loop, merged high-confidence findings, and an optional capped fix cycle you can trial safely with dry-run and safe-off commit defaults.
- Fingerprinted review findings with multi-source confidence boosts
- Optional bounded auto-fix iterations with loop-thrash protection
- Review runs controllable via dry-run and safe-off commit/PR-comment defaults
By the numbers
- 1 plugin in the marketplace (agent-review-pipeline)
- Plugin version 6.5.0
- 3 parallel review perspectives per iteration by default
Plugins in this marketplace
1 plugin - install individually after you add the marketplace.
Recommended Marketplaces
How it compares
Autonomous multi-engine review pipeline plugin, not a single static code-review skill or a general agent toolkit.
FAQ
Who is Agent Review Pipeline for?
Developers using Claude Code who want repeatable, multi-perspective code review with confidence scoring before they merge or release.
When should I use Agent Review Pipeline?
Use it in the ship phase when a change is ready for review, especially before opening or updating a PR or when you want an optional bounded auto-fix pass after review.
How do I add Agent Review Pipeline to my agent?
Add the onchainyaotoshi/Agent Review Pipeline marketplace to Claude Code, install the Agent Review Pipeline plugin from the catalog, configure Codex and Gemini access as required, and start with --dry-run before enabling commits or PR comments.