跳到正文
cccskills
免费在 GitHub 公开

ls-wf-planning-critic-loop

Use when creating decision-complete plans through grounding, capped clarification, subagent delegation, and critic iteration.

中文简介正在准备中。当前显示原文说明。

查看安装方法

包含的文件(2)

  • SKILL.md2.2 KB
  • workflow.yaml3.1 KB

SKILL.md(原文)

安装前可以查看交给 AI 代理的指令内容。

Use this workflow package for non-trivial planning work where the agent must produce a structured plan before implementation, handoff, or approval.

Ground first: inspect available repo files, docs, config, schemas, current GitHub state, or local system facts before asking the user. Do not ask questions that non-mutating inspection can answer.

Clarify only when it materially changes the plan. Ask at most three user questions total, one at a time, using the reverse decision-tree style from ls/docs/DECISION_TREE_WORKFLOW.md: A-D options, preferred choice, and rationale. Use available question/input/TUI tooling when supported; otherwise render clear markdown or plain-text choices. This three-question cap is local to this planning workflow and does not change the default ls-wf-spec-clarify-reverse protocol.

For non-trivial plans, use subagents when they reduce risk or context load: explorer for broad repo discovery, researcher for current external facts, tester for validation plans or long command logs, and reviewer as an independent critic when that materially improves evidence and is available and authorized. Keep the controller responsible for verification, checkpoints, and final plan quality. Do not force delegation for trivial direct tasks or when direct execution is the only useful route.

Review plans for requirement coverage, resolved decisions, grounding evidence, verification, safety, and scope. Use an independent critic when it materially improves evidence and is available and authorized. Repeat review after material corrections or when material findings remain; do not impose a minimum number of rounds or a numeric satisfaction quota. Turn unresolved prerequisites into explicit blockers, and keep the controller responsible for decision completeness. Submit the resulting plan through the active client's plan-submission mechanism; do not require literal <proposed_plan> output.

Primary references: ls/docs/DECISION_TREE_WORKFLOW.md, ls/docs/WORKFLOW_STANDARD.md, ls/docs/WORKFLOW_PACKAGES.md, and ls/docs/SKILLS_AND_RULES.md.

评价

还没有评价。欢迎分享你的使用感受。

评论

无需登录即可发表。欢迎分享使用感受、提问,或说明在哪些环境下无法运行。

还没有评论。来写第一条吧。

提交后会立即公开。请勿填写邮箱、电话等个人信息。URL 最多 2 个。

0 / 1000 字

此表单受 reCAPTCHA 保护,适用 Google 的隐私权政策和服务条款。

同一仓库的技能

Guide accessibility reviews for UI, documents, and workflows. Use for WCAG-oriented audits, keyboard review, semantics, contrast, and remediation planning.

中文简介正在准备中。当前显示原文说明。

CruxExperts/LocalSetup62026年10月21日 更新

Use when choosing a LocalSetup Agent-* lane for exploration, research, bounded implementation, or review. Consult the static capability matrix; do not probe account or network availability.

中文简介正在准备中。当前显示原文说明。

CruxExperts/LocalSetup62026年10月21日 更新

Guide agent email inbox workflows with prompt-injection defenses. Use for inbound email processing, webhooks, routing, and safety review.

中文简介正在准备中。当前显示原文说明。

CruxExperts/LocalSetup62026年10月21日 更新

Navigate and understand codebases using agentlens hierarchical documentation. Use when exploring new projects, finding modules, locating symbols in large files, finding TODOs/warnings, or understanding code structure.

中文简介正在准备中。当前显示原文说明。

CruxExperts/LocalSetup62026年10月21日 更新

Use for Agent Q signed encrypted file-drop or mail ship and ingest, private registry v2 setup, queue status, and transport maintenance.

中文简介正在准备中。当前显示原文说明。

CruxExperts/LocalSetup62026年10月21日 更新

Inventory wrapper for the alirezarezvani Claude skills bundle. Use when triaging its broad skill catalog and large script surface before any targeted import.

中文简介正在准备中。当前显示原文说明。

CruxExperts/LocalSetup62026年10月21日 更新

查看 CruxExperts 的全部技能

报告此技能的问题