Loopx — Development skill for Claude Code
LoopX cross-project quality gate kit for Codex and Claude Code.
How to install Loopx
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open rye567/loopx and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Loopx does
LoopX cross-project quality gate kit for Codex and Claude Code.
Alternatives in Development
- Stop Review Gate — You are reviewing uncommitted changes before the author finishes their turn 664 ★
- Project Memory: AgentSys — Cross-tool compatible: Works across all 5 platforms - Claude Code, OpenCode, Codex CLI, Cursor, and Kiro 622 ★
- Verify Theme — Run the full Graphite quality gate (drift, tokens, contrast, YAML) and report pass/fail 452 ★
README
LoopX
[](https://github.com/rye567/loopx/actions/workflows/ci.yml) [](LICENSE) [](https://github.com/rye567/loopx/releases)
Staged workflow checks for AI coding agents.
[中文说明](README.zh-CN.md)

flowchart LR
A["Init
env check"] --> B["Requirement
intake + interview"]
B -->|"confirm-stage
human confirmation"| C["Spec draft
+ review"]
C --> D["Mode select
LIGHT / STANDARD / FULL"]
D --> E["Solution design
+ review"]
E -->|"confirm-stage
human confirmation"| F["Test design
+ review"]
F --> G["Development
can-write unlocks here"]
G --> H["Quality audit
code review
test execution"]
H --> I["Health check
release readiness"]
I --> J["Final report
final check + close"]
LoopX is a Git-maintained skill package for Codex and Claude Code. It keeps AI coding work from jumping straight into implementation by guiding risky changes through requirement interview, spec drafting, human-confirmed reviews, test planning, implementation, and release checks.
Use it when an agent is about to touch cross-module behavior, API contracts, permissions, tenant boundaries, state machines, SQL/MQ flows, or anything where "looks done" is not good enough.
Why LoopX
AI coding agents are fast, but they often skip requirements, invent acceptance criteria, mark reviews as passed without evidence, or edit business logic before the plan has been checked. LoopX adds a local controller and a staged workflow so every meaningful change leaves auditable artifacts before code changes ship.
| Without LoopX | With LoopX |
|---|---|
| Agent starts coding from a vague prompt | Agent starts with requirement intake and interview |
| Acceptance criteria are i |
Related Skills
Do SOUL Alaya
The local-first memory plane for CLI coding agents. Codex and Claude Code attach via MCP; SQLite-backed, evide
Skill Creator Pro
A cross-framework skill creation tool with dual modes (quick-draft and guided-alignment) for creating high-qua
Quality Review
Run docs-reviewer, code-review, and diff-scoped E2E in parallel as a quality gate
Takahirom Claude Code Marketplace
Plugins such as Quality Gate and Skill Extractor
Gate Check
Execute the three mandatory quality gates for the current project.
Phase Review
Run post-execution quality gate - verifies all deliverables before marking a track complete
Related Agents
Ensemble Curator
High-volume triage agent for external-AI-agent PRs (Codex, Jules, Hermes, Droid, Aider, etc.). Analyzes, conso
Crucible Qg Judge
Stagnation judge for Crucible quality-gate — mechanical cross-round comparison of finding sets. Pinned to Sonn
Independent Reviewer
Product-technical alignment reviewer. Spawned in a tmux pane via OpenAI Codex CLI with zero inherited context.