Stop AI coding agents from guessing. MCP server + Claude Code plugin that enforces intent alignment BEFORE coding: PRD → intent-confidence gate → Mermaid contracts (state machines / sequence diagrams / decision tables) → mechanical lint with zero CRITICAL → then code. Human-in-the-loop, file-persisted, zero-config.
Skill Claude CodeCodex
Use when implementing or modifying code for a requirement (实现需求/编码/写代码/开工) that has an intent-gate contract (.harness/requests//summary.md exists) - an addendum layer ON TOP of your existing coding skills/workflow (superpowers, project CLAUDE.md, etc.): it never replaces them, it only adds "code is generated FROM the…
Skill Claude CodeCodex
Use when the user asks for 红蓝对抗 / 蓝军评审 / 对抗评审 / 需求评审 (adversarial review of a requirement analysis), when the user asks to 整改/按 findings 整改 (red-team revision of the analysis per review findings, §5.5), or after a complex requirement's summary.md is delivered and the human opts in - the blue army reviews red army…
Skill Claude CodeCodex
Use when analyzing a requirement/PRD (分析需求/解析需求/解析PRD/分析PRD), building state machines (画状态机), sequence diagrams (时序图), decision tables (决策表) or DDL from a spec, or resuming an interrupted requirement analysis (用户说"继续") - drives the three-level intent-alignment funnel (code evidence, registered inference, structured…
Skill Claude CodeCodex
Use when starting any conversation in a project that has the intent-gate MCP - establishes requirement intent-alignment discipline (never guess on ambiguity), where the analysis playbook lives, and the two optional add-ons (red-blue adversarial review skill, intent-gate-service sister service for DingTalk group…