zixun-github/ai-dlc · Archived

spec-context

Use when 需要在 dlc-dev 的 Spec 流程中定位当前 spec pack（FEATURE_DIR）、避?

First seen Jul 28, 2026

Installation

$ npx skills add zixun-github/ai-dlc --skill spec-context

View on skills.sh ↗ GitHub repository ↗

Stronger alternatives

This repository is archived — consider an actively maintained alternative.

subagent-driven-development Same source · more installs

zixun-github/ai-dlc

Use when executing implementation plans with independent tasks in the current session

spec-product-prd Same source · more installs

zixun-github/ai-dlc

Use when 需要在 dlc-dev 的产品需求 Spec 流程执行 R2，将 requirements/solution.md 转写为可交付、…

spec-product-clarify Same source · more installs

zixun-github/ai-dlc

Use when 在 dlc-dev 的 spec 分支上需要完成 R1（raw→solution）的需求澄?

spec-design-research Same source · more installs

zixun-github/ai-dlc

Use when 需要在 Spec 级设计阶段执行 D1 research（产出 `{FEATURE_DIR}/design/research.md`），或…

Similar popular skills

Related neighbors and high-traction skills in the same topics — useful to compare before installing.

context-map Similar name

github/awesome-copilot

Generate a map of all files relevant to a task before making changes

9.7K installs

what-context-needed Similar name

github/awesome-copilot

Ask Copilot what files it needs to see before answering a question

context7-cli Similar name

upstash/context7

Use the ctx7 CLI to fetch library documentation, manage AI coding skills, and configure Context…

6.2K installs

context7-mcp Similar name

upstash/context7

This skill should be used when the user asks about libraries, frameworks, API references, or ne…

4.8K installs

data-context-extractor Similar name

anthropics/knowledge-work-plugins

Generate or improve a company-specific data analysis skill by extracting tribal knowledge from …

2.8K installs

nemo-mbridge-perf-moe-long-context Similar name

Long-context MoE training guidance for Megatron Bridge. Covers CP sizing, selective recompute, …

1.8K installs

Also in this package

Other skills from zixun-github/ai-dlc · top by installs.

npx skills add zixun-github/ai-dlc

subagent-driven-development 5 spec-product-prd 5 spec-product-clarify 5 spec-design-research 5 spec-design 5 project-discover-preflight-scope 5 project-discover-modules-contracts 5 spec-receiving-code-review 4 spec-requesting-code-review 4 using-aidlc 4

Browse all from zixun-github/ai-dlc

More details

Agent compatibility

Declared targets from SKILL.md / docs. Unmarked agents are not listed — the skill may still install via the CLI.

Claude Code Not declared

Cursor Not declared

Codex Not declared

GitHub Copilot Not declared

Windsurf Not declared

Gemini CLI Not declared

Cline Not declared

OpenCode Not declared

Repository health

Stars 4

Default branch master

Open issues 1

Status Archived

Package contents

Files included with this skill beyond the listing page.

skill md SKILL.md 3,017 B
docs SUMMARY.md 228 B

History

Jul 28, 2026 First seen on skills.sh
Sep 9, 2026 First recorded snapshot · 4 installs

SKILL.md

Spec 上下文定位（FEATURE_DIR）

概览

读写 requirements/*.md 前必须先确定 {FEATURE_DIR}。以当前 Git 分支名为锚点，禁止猜路径。

分支格式：

Spec 分支：{num}-{short-name}（如 001-user-auth）→ FEATURE_DIR=.aidlc/specs/<分支名>。
任务分支（协作）：{num}-{short-name}/T1-task-slug（如 001-user-auth/T1-api）→ FEATUREDIR 仍指向 同一 Spec Pack（001-user-auth 目录），不指向任务子目录。脚本输出会带 TASKWORKBRANCH、TASKDIR、TASKPACKDIR（若 Task Pack 已存在）。

如果仓库包含 .gitmodules，spec-context 还应额外暴露 submodule 状态快照，供实现阶段校验分支一致性、detached HEAD 与脏工作区；但这不会改变 FEATURE_DIR 的解析规则。

开始时宣布：「我正在使用 spec-context 技能定位当前 Spec Pack（FEATURE_DIR）。」

何时使用

将读/写 {FEATURE_DIR}/requirements/*.md
不确定当前 spec pack 或出现分支/目录/.aidlc 相关上下文错误

做法：复用已有结果，否则执行脚本

优先复用：若本会话已成功回显过 FEATUREDIR=...，且该目录存在并含 requirements/，则直接复用并回显 FEATUREDIR=... (reuse)。

否则：拼接 <本SKILL.md目录>/scripts/ 下的脚本绝对路径并执行，从输出中读取 FEATURE_DIR=<path> 行。

对于实现阶段调用方（如 spec-plan、spec-execute），如果脚本还输出了 SUBMODULESETJSON=... 或等价字段，也应一并保留并传递给后续步骤使用。

PowerShell

powershell -NoProfile -ExecutionPolicy Bypass -File "<本SKILL.md目录>/scripts/spec-common.ps1" -SkillName "<caller-skill-name>"

Bash

"<本SKILL.md目录>/scripts/spec-common.sh" "<caller-skill-name>"

-SkillName / 第一参数：替换为当前执行的技能名（如 spec-plan）。
<本SKILL.md目录>：替换为本 SKILL.md 所在目录的绝对路径。

硬规则

读写 requirements/*.md 前必须先回显 FEATURE_DIR=...（复用或脚本二选一）
脚本失败则立即停止，不得继续写文件
后续路径一律以 FEATURE_DIR 值为前缀
若仓库存在 .gitmodules，实现阶段不得绕过脚本去“手猜” submodule 路径或分支状态

常见错误与 Red Flags

未传 -SkillName 或照抄 <caller-skill-name> 字面量
非 spec / 非任务分支（既不符合 {num}-{short-name}，也不符合 {num}-{short-name}/T1-slug）
手写 .aidlc/specs/... 路径；上下文失败后仍继续生成
把 submodule 分支当成新的 Spec 身份，试图用它单独推导 FEATURE_DIR

npx skills add zixun-github/ai-dlc --skill spec-context