Agent Skills › zereight/gitlab-mcp › plan

plan

GitHub

通过智能交互生成全面可执行的工作计划，支持访谈、直接生成、共识评审和现有计划审查模式。

.github/skills/plan/SKILL.md zereight/gitlab-mcp

触发场景

用户说 plan this, plan the, let's plan, make a plan, how should we approach 需要制定项目或功能实施策略时

安装

npx skills add zereight/gitlab-mcp --skill plan -g -y

更多选项

非标准路径

npx skills add https://github.com/zereight/gitlab-mcp/tree/main/.github/skills/plan -g -y

不安装直接使用

npx skills use zereight/gitlab-mcp@plan

指定 Agent (Claude Code)

npx skills add zereight/gitlab-mcp --skill plan -a claude-code -g -y

安装 repo 全部 skill

npx skills add zereight/gitlab-mcp --all -g -y

预览 repo 内 skill

npx skills add zereight/gitlab-mcp --list

SKILL.md

Frontmatter

{
    "name": "plan",
    "description": "Strategic planning with optional interview workflow. Activate when user says: plan this, plan the, let's plan, make a plan, how should we approach.\n",
    "argument-hint": "[--direct|--consensus|--review] <task description>"
}

Plan

Creates comprehensive, actionable work plans through intelligent interaction. Auto-detects whether to interview (broad requests) or plan directly (detailed requests).

Modes

Mode	Trigger	Behavior
Interview	Default for broad requests	Interactive requirements gathering
Direct	`--direct`, or detailed request	Skip interview, generate plan directly
Consensus	`--consensus`, "ralplan"	Planner → Architect → Critic loop
Review	`--review`	Critic evaluation of existing plan

Interview Mode (broad/vague requests)

Classify request: broad triggers interview
Ask ONE focused question at a time for preferences, scope, constraints
Gather codebase facts via @explore BEFORE asking user
Consult @analyst for hidden requirements
Create plan when user signals readiness

Direct Mode (detailed requests)

Optional brief @analyst consultation
Generate comprehensive work plan immediately

Consensus Mode (`--consensus` / "ralplan")

@planner creates initial plan with RALPLAN-DR summary (Principles, Decision Drivers, Options)
@architect reviews for architectural soundness (sequential, NOT parallel with critic)
@critic evaluates quality criteria (after architect completes)
Re-review loop (max 5 iterations) if critic rejects
Apply improvements on approval
Final plan includes ADR (Decision, Drivers, Alternatives, Why chosen, Consequences)