deps

分类: 数据与AI | 上传者: ivanviragine | 下载: 0 | 版本: v1.0（最新）

Analyze a set of GitHub issues and identify dependency relationships between them. Reads issue titles and bodies, determines execution order, and outputs a structured dependency file that wade uses to generate Mermaid diagrams and update issue bodies. Use when multiple related issues need ordering. --- # Dependency Analysis Analyze a set of GitHub issues and determine the dependency relationships between them. Output a structured file that `wade` will use to generate dependency graphs and update issue bodies. ## When to activate - After `wade plan-task` creates multiple issues - When `wade task deps` is run on existing issues - When the user asks to analyze dependencies between issues > **Note:** `wade task deps` first attempts headless analysis (AI tools that > support `--print`/`--prompt`). If headless fails, it falls back to interactive > mode: passes the analysis prompt directly to the AI tool as an initial message, > then reads the output from a file after exit. ## Input

更新日志: Source: GitHub https://github.com/ivanviragine/wade

目录结构

当前层级: templates/skills/deps/

📄 SKILL.md 3.4 KB

SKILL.md

---
name: deps
description: >
  Analyze a set of GitHub issues and identify dependency relationships between
  them. Reads issue titles and bodies, determines execution order, and outputs
  a structured dependency file that wade uses to generate Mermaid diagrams
  and update issue bodies. Use when multiple related issues need ordering.
---

# Dependency Analysis

Analyze a set of GitHub issues and determine the dependency relationships
between them. Output a structured file that `wade` will use to generate
dependency graphs and update issue bodies.

## When to activate

- After `wade plan-task` creates multiple issues
- When `wade task deps` is run on existing issues
- When the user asks to analyze dependencies between issues

> **Note:** `wade task deps` first attempts headless analysis (AI tools that
> support `--print`/`--prompt`). If headless fails, it falls back to interactive
> mode: passes the analysis prompt directly to the AI tool as an initial message,
> then reads the output from a file after exit.

## Input

You will receive a context file containing one or more issues in this format:

```
=== Issue #41: Add user preferences schema ===
<issue body text>

=== Issue #42: Add preferences API endpoints ===
<issue body text>

=== Issue #43: Add preferences UI panel ===
<issue body text>
```

## Step 1: Analyze relationships

For each pair of issues, determine if there is a dependency. A dependency
exists when:

- One issue creates something (schema, module, API) that another issue uses
- One issue sets up infrastructure that another issue builds upon
- The implementation order matters for correctness (not just convenience)

**Do NOT create dependencies for:**
- Soft preferences ("it would be nice to do A first")
- Shared concerns without actual data/code flow
- Testing dependencies (tests can usually be written independently)

## Step 2: Write the deps file

Write the dependency file to the path specified in your instructions.
Use this exact format:

```
# Dependencies (A -> B means "A must be done before B")
41 -> 42   # schema must exist before API can use it
41 -> 43   # schema must exist before UI can display it
42 -> 43   # API endpoints needed for UI to call
```

Rules:
- One edge per line: `<number> -> <number>`
- Comments after `#` explaining the reasoning (required for each edge)
- Only use issue numbers from the input context
- The graph must be a DAG (no cycles)
- If no dependencies exist, write only: `# No dependencies found`

> **Note:** In the Mermaid diagram that wade generates from this file, arrows are
> rendered in the conventional DAG direction: **dependent → dependency** (e.g., `42 --> 41`
> means issue 42 depends on 41). This is the reverse of the `A -> B` notation above, which
> is only the internal file format used to communicate with wade.

## Step 3: Confirm completion

After writing the file, tell the user:
> "Dependency file written to `<path>`. Exit this session so wade can apply
> the dependencies to the issues."

## Rules

- **Only create edges for genuine technical dependencies** — not stylistic preferences.
- **Every edge needs a comment** explaining why the dependency exists.
- **No cycles** — if A depends on B, B cannot depend on A (directly or transitively).
- **Minimize edges** — if A→B and B→C, don't add A→C unless there's an independent reason.
- **Use only issue numbers from the input** — never reference issues not in the context file.

登录后下载/点赞/收藏 ❤ 3 | ★ 0

deps

目录结构

SKILL.md

举报内容

提示