Evidence snapshot reviewed Aug 30, 2026GitHub checked Aug 21, 2026
DSHub.
Plugin RSS
Source-reviewedStandalone SkillAutomation & Agents

~auto

Selects and runs the appropriate clarification, planning, build, QA, or product workflow until delivery or a real blocker.

Open source project ↗View on GitHub ↗More Skills

At a glance

What it does

Selects and runs the appropriate clarification, planning, build, QA, or product workflow until delivery or a real blocker.

Capabilities
Automation & AgentsWorkflow AutomationAgentsTesting

Before you choose it

Routes work by scope and risk, follows the current project workflow when available, and continues across planning, implementation, QA, and closeout instead of stopping at intermediate handoffs.

Best for

HelloAGENTS users who want a task carried through the appropriate workflow with fewer manual handoffs.

Common tasks

  • Route an unfamiliar task
  • Continue a multi-stage implementation
  • Carry completed work through QA and closeout

Permissions and data

The skill guides a local workflow; external behavior was not independently audited.

Permissions
  • Uses the declared plugin or skill workflow.
Data handling
  • Data handling was not independently audited.
External services
  • External services follow the supplied project documentation.
Credentials
  • Credential requirements were not independently tested.

Limitations

  • Authorization boundaries still apply to external side effects and high-risk actions
  • The command stops when required information or authority is missing

What DSHub checked

  • The pinned source and supplied artifact evidence were captured.

What DSHub did not check

  • Installation and runtime behavior were not tested.

Pinned install

Primary action

This standalone skill does not have a DSH Plugin install action. Use its source documentation for the delivery method.

Visit the source project ↗

Maintainer source

Skill instructions

View at commit 79c0b48 ↗
Maintainer-authored contentCaptured from skills/commands/auto/SKILL.md on Aug 30, 2026. The text and repository-relative media are fixed to commit 79c0b482fd20 with content hash 2c4c06f6b1db; provider-hosted badges may update independently. SKILL.md commands are upstream documentation; use the type-correct primary action above and verify it against this pinned source.

name: ~auto description: 自动执行命令 — 自动选择并依次执行 ~ask / ~plan / ~build / ~qa / ~prd，默认持续推进直到交付完成（~auto 命令） policy: allow_implicit_invocation: false

Trigger: ~auto <任务描述>

~auto 是自动执行命令。它根据任务类型、复杂度、风险等级与项目状态，在 ~ask、~plan、~build、~qa、~prd 之间选择合适主路径，并连续推进。 ~auto 不止做一次选路；主路径一旦确定，就按需要继续执行后续阶段，默认持续推进直到完成交付，只有命中 HelloAGENTS 阻塞判定时才停下。

铁律

  • 不为了“自动化”而强行走重流程
  • 复杂度与风险不足以支撑更重路径时，优先选更轻但能保证质量的路线
  • T3 高风险或不可逆操作默认不直接进入 ~build；优先先走 ~plan 或 ~prd，纯质量审查、验真或收尾请求才可先进入 ~qa
  • 主路径一旦确定，立即读取对应 command skill，并在阶段完成后继续执行后续阶段，避免同一任务重复探索或重复等待
  • 选路不替代授权；涉及外部副作用或高风险不可逆操作时，仍遵守 HelloAGENTS 阻塞判定与确认规则
  • 用户显式使用 ~auto，表示已授权在当前任务边界内沿选定主路径持续执行；若当前运行在 Codex /goal 下，/goal 只提供长程续跑与预算，~auto 仍按方案包、state_path 与验证契约推进；~plan / ~prd 作为中间阶段时，不再额外询问“是否开始执行”，除非仍有真实阻塞；不得把 🔄 下一步 当作阶段交接或继续执行占位
  • 优先消费当前上下文中已注入的“选路与分层”结果、当前工作流约束与项目状态；不要在 ~auto 内另建一套关键词路由表

流程

0. 当前工作流优先

  • 若当前上下文中已注入“当前工作流提示”或“当前工作流约束”，优先服从其中的推荐下一命令 / 主路径
  • 默认原则：
    • 活跃方案包不完整或缺少任务清单 → 先 ~plan
    • 活跃方案包仍在执行 → 先 ~build，完成当前实现后再 ~qa
    • 活跃方案包已闭合 → 先 ~qa 或收尾
    • Codex active goal 已关联方案包 → 按 tasks.md 未完成 AFK 项、contract.json 与 state_path 确定主路径，不把 goal 目标原文替代方案包
  • 只有当用户明确要求换方向、重做方案，或现有方案已失效时，才偏离当前推荐重新规划

1. 选路

  • 先按当前上下文里已注入的“选路与分层”语义约束判断，不依赖关键词命中做机械分流
  • 若当前上下文没有足够的注入约束，再结合以下信号补足判断：影响范围、风险等级、是否需要结构化产物、是否已有活跃方案包、用户是否只想先比较方向
  • 选路优先级：
    • 方向不明 / 价值与范围仍不清楚 / 点子与方向比较 → ~ask
    • 明确要求质量审查 / 验真 / 跑检查 / 收尾 → ~qa
    • 0 到 1 / 产品级 / 多维规格 → ~prd
    • 多文件功能 / 架构变更 / 新项目规划 → ~plan
    • 明确修复 / 小范围实现 / 按现有方案实现 → ~build

2. 按 Tier 校正

  • T0 → 走 ~ask，不创建项目文件
  • T1 → 在 ~build / ~qa 间选择最短可交付路径
  • T2 → 需要结构化产物或范围未完全明确时优先 ~plan
  • T3 → 纯质量审查/验真走 ~qa；其余默认 ~plan 或 ~prd，待方案与风险边界明确后再进入实现

3. 读取对应命令并执行主路径

  • 选中 ask → 读取 skills/commands/ask/SKILL.md
  • 选中 plan → 读取 skills/commands/plan/SKILL.md
  • 选中 build → 读取 skills/commands/build/SKILL.md
  • 选中 qa → 读取 skills/commands/qa/SKILL.md
  • 选中 prd → 读取 skills/commands/prd/SKILL.md

不要额外读取未选中的 command skill。

4. 持续执行直到完成

  • 若主路径是 ~build → 完成实现后继续进入 ~qa，再进入收尾
  • 若主路径是 ~plan → 方案包写入后，若当前任务来自 ~auto 且未命中阻塞判定，直接继续进入 ~build，不要把“方案已形成”当作最终停点
  • 若主路径是 ~prd → PRD / 任务 / 契约写入后，若当前任务来自 ~auto 且未命中阻塞判定，按当前结果继续进入 ~build，必要时先补一轮轻量 ~plan
  • 若主路径是 ~qa → 完成质量闭环 / 收尾后结束
  • 若主路径是 ~ask，且用户本意就是探索/厘清，则在交互澄清输出后结束；若澄清后已有明确方向且当前任务仍要求写文件或改代码，则继续进入 ~plan 或 ~build
  • 若 Codex active goal 的目标已满足 → 仍先完成 ~qa 与 HelloAGENTS 收尾，再标记 goal complete；未满足时继续下一项可执行 AFK 任务

5. 何时允许停下

  • 仅在 HelloAGENTS 阻塞判定成立时停下：真实歧义、缺必需信息/文件/凭据、未授权外部副作用、高风险或不可逆操作
  • 不得把 ~plan / ~prd 中“是否进入执行”的默认询问原样带入 ~auto 流程
  • 不得把“给出方案”“给出任务列表”“未执行修改”“等待下一步确认”“给出建议下一步”当作 ~auto 的默认完成态；用户显式要求 ~auto 时，默认目标是把当前任务推进到可交付完成

Operate deliberately

Install and manage

Prerequisites and target Profile

Target： No native DSH Profile target.

Delivery： Skill Files — https://github.com/hellowind777/helloagents。

Compatibility and access

Requires_host_review： Not established by supplied evidence。

Review compatibility evidence ↗

Risk facts

Execution

Review instructions before use.

Evidence ↗
Evidence and editorial reviewManifest, Bundle patch, distribution and freshness

Immutable evidence

Review status and source activity

Human approved

Approved for publication after reviewing the source-linked content and immutable release record. AI assisted with the draft; the publication decision was human.

Human reviewed Aug 30, 2026, 10:05 AM UTC。GitHub facts last checked Aug 30, 2026, 9:53 AM UTC。

No material source change has been recorded since this evidence baseline.

Next step

Compare ecosystem artifact types →

Subscribe to material changes for ~auto →