AgentSkillsCN

skill-product-backlog-prioritization

运用 WSJF 方法管理并优先排序产品待办事项列表。

SKILL.md
--- frontmatter
name: skill-product-backlog-prioritization
description: Guidelines for managing and prioritizing the Product Backlog using WSJF.
tier: 2
version: 1.0

Backlog Prioritization (WSJF)

1. Objective

To prioritize features scientifically using Weighted Shortest Job First (WSJF).

2. Logic Locker (CRITICAL)

[!IMPORTANT] FORBIDDEN ACTION: You are strictly forbidden from calculating WSJF scores manually or mentally. Humans and LLMs are bad at floating point arithmetic. Use the script.

3. Core Tooling

You MUST use the provided Python script to calculate scores and sort the backlog.

How to Prioritize

Run the following command:

bash
python3 [skill_path]/scripts/calculate_wsjf.py --file docs/PRODUCT_BACKLOG.md

Note: [skill_path] is the path to this skill (e.g. .agent/skills/skill-product-backlog-prioritization).

4. Scoring Components

When adding a new item to the Backlog, you must estimate:

  1. •User Value (UV): Relative value to the customer (1-10).
  2. •Time Criticality (TC): Is there a deadline? (1-10).
  3. •Risk Reduction (RR): Does this open new options or fix security? (1-10).
  4. •Job Size (JS): Estimate of effort. Support "T-Shirt Sizes" (XS, S, M, L, XL) or relative fibonacci (1, 2, 3, 5, 8, 13, 20).
    • •XS / ~1d -> 1
    • •S / ~2-3d -> 2
    • •M / ~1w -> 5
    • •L / ~1m -> 13
    • •XL / ~2m+ -> 20

Formula (Handled by Script): WSJF = (UV + TC + RR) / Job Size

5. Artifact Standards

See examples/backlog_table_example.md for the correct table format. The table MUST include all 4 input columns and the WSJF output column.