AgentSkillsCN

prd-status

以状态、质量评估及元数据呈现所有 PRD 的仪表板视图。

SKILL.md
--- frontmatter
name: prd-status
description: Show a dashboard of all PRDs with status, quality assessment, and metadata
argument-hint:
user-invocable: true
disable-model-invocation: true
allowed-tools: [Read, Glob, Grep]

User Input

text
$ARGUMENTS

You MUST consider the user input before proceeding (if not empty).

Goal

Scan the prd/ directory and produce a dashboard report showing all PRDs with their status, metadata, section coverage, and a quick quality assessment. Output is displayed in the terminal only — no files are modified.

Execution Steps

1. Scan PRD Directory

  • •Use Glob to find all prd/*/prd.md files.
  • •If no PRDs are found, report: "No PRDs found in prd/ directory."
  • •If $ARGUMENTS is not empty, treat it as a filter (e.g., draft filters to PRDs with Draft status, or a feature name substring).

2. Extract Metadata per PRD

For each PRD file found, read and extract:

  1. •Feature name — From the parent directory name
  2. •Title — From the first # heading
  3. •Status — From metadata section (Draft / In Review / Approved / Active Development)
  4. •Version — From metadata section
  5. •Last updated — From metadata section
  6. •Author — From metadata section
  7. •Word count — Total words in the document
  8. •Section count — Number of ## headings that map to the 22-section template
  9. •Required sections present — Count of the 12 required sections found (out of 12)
  10. •Has context file — Check if prd/<name>/prd-context-*.md exists
  11. •Has summary — Check if prd/<name>/prd-summary.md exists
  12. •Has tasks — Check if prd/<name>/tasks.md exists
  13. •Placeholder count — Count of [NEEDS INPUT], TODO, TBD, ??? markers

3. Quick Quality Assessment

For each PRD, assign a quick health indicator based on:

  • •Ready — All 12 required sections present, no placeholders
  • •Good — 10-11 required sections present, ≤ 2 placeholders
  • •Needs Work — 7-9 required sections present, or > 2 placeholders
  • •Incomplete — < 7 required sections present

This is NOT a full analysis (use /prd-analyze for that). It's a fast heuristic.

4. Produce Dashboard Report

Output a Markdown report with the following structure:

code
## PRD Dashboard

**Date**: <current date>
**PRDs found**: X

---

### Overview

| PRD | Status | Version | Health | Sections | Words | Updated |
|-----|--------|---------|--------|----------|-------|---------|
| <name> | Draft | 1.0 | Ready | 12/12 | 1,329 | 2026-01-31 |
| <name> | Active | 2.0 | Good | 11/12 | 749 | 2026-01-23 |

---

### Artifacts

| PRD | Context File | Summary | Tasks | Placeholders |
|-----|-------------|---------|-------|-------------|
| <name> | Yes | No | No | 0 |

---

### Quick Actions

- PRDs needing attention: <list any with "Needs Work" or "Incomplete" health>
- PRDs without summaries: <list> → Run `/prd-export <path>`
- PRDs without tasks: <list> → Run `/prd-tasks <path>`
- PRDs with placeholders: <list> → Run `/prd-refine <path>`

5. Optional Filtering

If $ARGUMENTS is provided:

  • •Filter by status keyword (e.g., draft, approved)
  • •Filter by feature name substring (e.g., todo)
  • •Show only matching PRDs in the dashboard

Operating Principles

  • •NEVER modify files — this is strictly read-only
  • •Keep the scan fast — read only the first 100 lines of each PRD for metadata extraction
  • •Use Grep for counting placeholders (faster than reading entire files)
  • •If a metadata field is missing, show — in the table
  • •Sort PRDs by last updated date (most recent first)