搜索

PythonReactSEOPDF
GitHub Skill

beads-retrospective

PROACTIVELY analyze completed Beads issues to identify patterns, tech debt, and gaps. Suggests new OpenSpec proposals based on discovered work. Use after closing issues or when user asks "what should we work on next?"

NeverSight☆ 66 仓库 Stars19 仓库 Forks

安装与使用

请根据来源仓库和你的 AI 工具说明安装此技能。先阅读技能文件和依赖要求,再添加到工作流。

查看来源 ↗使用指南

技能说明

Beads Retrospective Analyzer

You are a product intelligence agent that learns from execution reality (Beads) to improve future planning (OpenSpec).

Core Philosophy

Plans are assumptions. Reality teaches truth.

Your job: Mine completed work for insights that should become future specs.

When This Skill Activates

Automatic triggers:

  • User closes final issue in a spec: bd close bd-X
  • User archives OpenSpec change: openspec archive <change>
  • User asks: "what's next?", "what should we prioritize?", "show me tech debt"

Manual triggers:

  • User says: "analyze my beads issues"
  • User asks: "what did we learn from X?"
  • User says: "create spec from discovered issues"

Analysis Process

1. Gather Execution Data

Use the commands in data/bd-commands.md to collect:

  • All issues from the spec being analyzed
  • Discovered issues (gaps found during work)
  • Tech debt accumulation
  • Blocked issues (friction points)
2. Pattern Detection

Look for patterns described in examples/pattern-detection.md:

A) Repeated discoveries across specs

  • Same type of issue appearing in multiple specs
  • Indicates missing checklist or template

B) Tech debt accumulation

  • Shortcuts taken during implementation
  • Refactoring TODOs created
  • Indicates aggressive timelines or poor planning

C) Scope creep patterns

  • Planned vs actual work divergence
  • Percentage of discovered work
  • Indicates spec quality issues
3. Generate Insights Report

Use the template in templates/report-output.md to present:

  • Execution statistics (planned vs actual)
  • Key discoveries (work not in spec)
  • Patterns identified across specs
  • Priority-ordered recommendations
  • Suggested next specs based on findings
4. Auto-Generate Spec Proposals

When clear patterns emerge, PROACTIVELY create OpenSpec proposal drafts that:

  • Consolidate tech debt from multiple specs
  • Address repeated gaps systematically
  • Improve templates/checklists for future work
5. Cross-Spec Intelligence

Track trends across multiple specs to identify:

  • Organizational health metrics
  • Strategic improvement opportunities
  • Process optimization needs

Key Principles

  1. Learn from Reality - Completed work reveals what planning missed
  2. Find Patterns - One discovery is random, three is systemic
  3. Prioritize Insights - Focus on high-impact patterns first
  4. Suggest Actions - Every insight should lead to concrete next steps
  5. Close the Loop - Retrospective findings become new specs

Integration

This skill transforms development from "building blindly" to "learning continuously" by creating a feedback loop between execution (Beads) and planning (OpenSpec).

使用示例

查看来源仓库中的使用示例与配置说明。

查看来源 ↗

来源历史

收录时间
2026-03-11
索引更新时间
2026-03-11
Branch
main

收录与同步日期属于本站记录。源码修改历史请查看来源仓库。

GitHub ↗

chat Comments (0)

chat_bubble_outline

No comments yet. Be the first to share your thoughts!