Token导航 LogoToken导航TokenDH.com
开发只读github未标认证来源可访问许可证需确认审计通过

spec-driven-propose规范驱动的建议

Agent Skill

spec-driven-propose 用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要围绕仓库状态、代码变更或协作事项进行整理时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

832

周安装

35

GitHub Stars

公开资料未说明

下载量

291
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:spec-driven-propose(规范驱动的建议)
来源仓库:https://github.com/kw12121212/auto-spec-driven
仓库路径:skills/spec-driven-propose
安装命令:
npx skills add https://github.com/kw12121212/auto-spec-driven --skill spec-driven-propose
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/kw12121212/auto-spec-driven --skill spec-driven-propose

简介

用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息,适合在 Codex、Claude、Cursor、Gemini CLI 中围绕仓库状态、代码变更或协作事项进行整理。

  • 可结合来源仓库、安装命令和原始 README 继续核验具体用法。
  • 安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。
  • 通过 npx skills add 命令从指定 GitHub 仓库安装 spec-driven-propose 技能。
  • 当前无额外说明,需参考原始 SKILL.md 获取完整功能细节。

SKILL.md

You are helping the user create a new spec-driven change proposal.

Do not ask follow-up questions or require confirmation during the proposal stage. Derive the strongest proposal you can from the user request, project context, and existing specs. Record any remaining ambiguity in questions.md for /spec-driven-apply to surface before implementation begins.

This Skill's Commands

If you cannot remember the exact command used by this skill, look it up here before running anything. Do not guess.

propose: node {{SKILL_DIR}}/scripts/spec-driven.js propose <name>
verify: node {{SKILL_DIR}}/scripts/spec-driven.js verify <name>

Prerequisites

The .spec-driven/ directory must exist at the project root. Before proceeding, verify:

ls .spec-driven/

If this fails, the project is not initialized. Run /spec-driven-init first.

Steps

  1. Determine the change name — use the user-provided kebab-case name when one is already available. Otherwise, derive a short kebab-case name from the request and the proposal scope yourself instead of stopping to ask for one.
  2. Read project context and existing specs — read the following before generating anything:

- .spec-driven/config.yaml — use context to inform content; treat rules as binding constraints; note any fileMatch entries that apply to files this change will touch - .spec-driven/specs/INDEX.md — identifies all existing spec files and their scope - Every spec file referenced in INDEX.md that this change is likely to touch — read the full content to understand existing requirements before writing MODIFIED or ADDED entries

  1. Scaffold the change — run: node {{SKILL_DIR}}/scripts/spec-driven.js propose <name> This creates .spec-driven/changes/<name>/ with seeded artifact templates.
  2. Fill proposal.md — write a clear, concise proposal covering:

- What: What the change does (observable behavior, not implementation) - Why: Motivation and context - Scope: What is in scope, what is explicitly out of scope - Unchanged Behavior: List existing behaviors that must not break — things adjacent to or potentially affected by this change. Leave blank if truly nothing is at risk.

  1. Fill design.md — write the technical approach:

- Approach: How you'll implement it at a high level - Key Decisions: Significant choices and their rationale - Alternatives Considered: What was ruled out and why

  1. Populate specs/ delta files — look at the project's .spec-driven/specs/ directory structure. For each spec file that this change touches, create a corresponding file under .spec-driven/changes/<name>/specs/ mirroring the same relative path (e.g. specs/auth/login.mdchanges/<name>/specs/auth/login.md). If the change introduces a new spec area, create the new relative path that should exist under .spec-driven/specs/ after archive. Mirror the main spec path exactly. For example, .spec-driven/specs/skills/planning.md maps to .spec-driven/changes/<name>/specs/skills/planning.md. Use this canonical sample as the format target: --- mapping: implementation: - path/to/implementation.ts tests: - test/path/to/test.ts --- ## ADDED Requirements ### Requirement: new-capability The system MUST provide <observable behavior>. #### Scenario: success - GIVEN <precondition> - WHEN <action> - THEN <result> ## MODIFIED Requirements ### Requirement: existing-capability Previously: The system MUST <old behavior>. The system MUST <new behavior>. ## REMOVED Requirements ### Requirement: old-capability Reason: This behavior is removed because <reason>. Each delta file uses ADDED/MODIFIED/REMOVED sections with the standard format:

- YAML frontmatter with mapping.implementation and mapping.tests when related files are knowable from repository context - ### Requirement: <name> headings and RFC 2119 keywords (MUST/SHOULD/MAY) - #### Scenario: blocks (GIVEN/WHEN/THEN) where helpful - ADDED: new requirements; MODIFIED: changed requirements (include Previously: note); REMOVED: removed requirements (include reason) - Omit sections that don't apply — do not leave empty sections - If this change has no observable spec impact, leave changes/<name>/specs/ empty — do not create a prose-only delta file that breaks the delta spec format - Do not invent mapping paths when the related implementation or test files are not clear; leave mapping completion to /spec-driven-apply

  1. Fill tasks.md — write a concrete implementation checklist using this canonical structure: ` # Tasks: <change-name> ## Implementation - [] Describe the first atomic implementation task - [] Describe the second atomic implementation task ## Testing - [] Run npm run lint — lint or validation task - [] Run npm test — unit test task ## Verification - [] Verify implementation matches proposal scope Each task uses - [] checkboxes. Tasks should be independently completable. The ## Testing section MUST satisfy these verification keyword requirements: If the relevant command cannot be determined confidently from repository context, add an open question to questions.md` instead of guessing. Do NOT add an "Update specs" task — the specs/ directory contains the spec artifacts.

- At least one task MUST contain a lint/validation keyword: lint, validate, validation, typecheck, type-check, or build - At least one task MUST contain a unit test keyword: unit test or unit tests - Both tasks MUST name an explicit runnable command in backticks or use a known runner (npm, pnpm, yarn, bun, node, bash, sh, pytest, jest, vitest, go, cargo, make, uv, poetry) - Paraphrasing these keywords (e.g. "run tests" instead of "run unit tests") will cause verify to fail

  1. Fill questions.md — document any open questions or ambiguities:

- For every unclear point (motivation, scope boundaries, technical approach, etc.), add an entry under ## Open: - [] Q: <specific question> Context: <why this matters / what depends on the answer> - If everything is clear, leave ## Open empty with a note: <!-- No open questions --> - Do NOT use [NEEDS CLARIFICATION] inline markers in any artifact — questions.md is the single place for all open questions

  1. Validate artifact format — run: node {{SKILL_DIR}}/scripts/spec-driven.js verify <name> Treat this as a final structure and format check before presenting the proposal.

- If it reports missing artifacts, malformed delta spec sections/headings, or unfilled placeholders you can safely fix, fix them immediately and rerun verify - If verify still reports a format or structure problem you cannot confidently repair, stop and report the issue to the user - If the only remaining error is open questions in questions.md, treat that as expected at proposal time and surface those questions clearly to the user

  1. Hand off — show the user the five files and summarize the scope, key decisions, and any open questions. Do not require a confirmation checkpoint after writing the proposal. If questions.md has open questions, make clear that /spec-driven-apply will surface them as an implementation blocker and require explicit user resolution before coding starts, and mention /spec-driven-modify only as an optional revision path.

Rules

  • Do not implement anything — this is planning only
  • Keep tasks atomic and verifiable
  • Do not ask follow-up questions or require a proposal-stage confirmation checkpoint
  • proposal.md describes *what and why*; design.md describes *how*; tasks.md is the checklist; questions.md is for open questions
  • Document ambiguities in questions.md — never guess at unclear requirements, and never use [NEEDS CLARIFICATION] inline markers
  • Do not add a post-proposal confirmation gate once the artifacts are written
  • Open questions are allowed at proposal handoff time; leave them in questions.md for /spec-driven-apply to surface and block on
  • Before finishing, rerun verify until all repairable format issues are fixed; if any non-question error remains, report it to the user
  • Keep spec-code mappings in frontmatter, not in requirement prose

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

33.48%
按下载量换算97

Claude

28.18%
按下载量换算82

Cursor

20.33%
按下载量换算59

Gemini CLI

9.77%
按下载量换算28

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

只读

该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills