review-orchestrator
Get multiple perspectives on your work — coordinate reviews across cognitive modes
Install via CLI (Recommended)
clawhub install openclaw/skills/skills/leegitw/review-orchestratorreview-orchestrator (審査)
Unified skill for selecting review types, spawning multi-perspective and cognitive review agents, and managing quality gates. Consolidates 5 granular skills into a single review system.
Trigger: レビュー要求 (review requested)
Source skills: twin-review, cognitive-review, review-selector, staged-quality-gate, prompt-normalizer
Installation
openclaw install leegitw/review-orchestrator
Dependencies:
leegitw/failure-memory(for context)leegitw/context-verifier(for file verification)
# Install with dependencies
openclaw install leegitw/context-verifier
openclaw install leegitw/failure-memory
openclaw install leegitw/review-orchestrator
Standalone usage: Review orchestration works independently for multi-perspective reviews. Integration with failure-memory enables automatic observation recording from review findings.
Data handling: This skill operates within your agent's trust boundary. When triggered,
it uses your agent's configured model for multi-perspective review orchestration. No external APIs
or third-party services are called. Review results are written to docs/reviews/ in your workspace.
What This Solves
One perspective has blind spots. This skill coordinates multiple review perspectives to catch what single-pass review misses:
- Twin review — technical and creative perspectives for balance
- Cognitive modes — analyzer ("what conflicts"), architect ("how to restructure"), implementer ("how to implement")
The insight: N=2 catches more than N=1. Different perspectives see different things. Coordinate them systematically.
Note: "Cognitive modes" are review perspectives with different analytical focus, not external API calls. Mode names (analyzer, architect, implementer) describe the review approach, not specific AI models or services.
Usage
/ro <sub-command> [arguments]
Sub-Commands
| Command | CJK | Logic | Trigger |
|---|---|---|---|
/ro select | 選択 | context×risk→type∈{twin,cognitive,code} | Explicit |
/ro twin | 双子 | spawn(technical,creative)→findings[] | Explicit |
/ro cognitive | 認知 | spawn(modes[])→analysis[] | Explicit |
/ro multi | 双視 | alias for /ro twin (multi-perspective review) | Explicit |
/ro gate | 門番 | staged_work→pass✓∨block✗ | Explicit |
Arguments
/ro select
| Argument | Required | Description |
|---|---|---|
| context | Yes | Description of work to review |
| --risk | No | Risk level: low, medium, high (auto-detected if omitted) |
/ro twin
| Argument | Required | Description |
|---|---|---|
| target | Yes | File path(s) or topic to review |
| --technical-only | No | Skip creative perspective |
| --creative-only | No | Skip technical perspective |
/ro cognitive
Metadata
Not sure this is the right skill?
Describe what you want to build — we'll match you to the best skill from 16,000+ options.
Find the right skillPaste this into your clawhub.json to enable this plugin.
{
"plugins": {
"official-leegitw-review-orchestrator": {
"enabled": true,
"auto_update": true
}
}
}Tags
Related Skills
q-kdb-code-review
AI-powered code review for Q/kdb+ — catch bugs in the most terse language in finance
retake-tv-agent
Go live on retake.tv — the livestreaming platform built for AI agents. Register once, stream via RTMP, interact with viewers in real time, and build an audience. Use when an agent needs to livestream, engage chat, or manage its retake.tv presence.
codex-review
Three-tier code quality defense: L1 quick scan, L2 deep audit (via bug-audit), L3 cross-validation with adversarial testing. 三级代码质量防线。
qa-check
Mandatory quality assurance for all dev work before publishing. Use BEFORE deploying any project to production. Validates build, tests browser functionality, checks mobile responsiveness, and ensures no broken links/images.
decision-log
AI-powered decision journal for solopreneurs — capture decisions with context, rationale, and expected outcomes, then review them later to learn from what you got right and wrong.