prove-your-worth

Ruthlessly audit project features for justification. Challenge every feature to prove its value with evidence or face removal. Uses MCP tools for research.

跳到安装

仓库: majiayu000/claude-skill-registry
最近来源活动: 2026年6月23日 12:15
检测到的 SKILL.md 语言: 英语
星标: 543
分支: 85

安装方式

默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。

用 Codex 或 Claude 帮你安装复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。

直接命令不会经过审查 Prompt；运行前请先检查来源。

npx skills add https://github.com/majiayu000/claude-skill-registry --skill prove-your-worth

命令会保持在同一行。复制前请横向滚动并检查完整内容。

想先保存到本地？可下载 SkillsMP 当前能够提供的文件。

检查来源文件

决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。

文件资源管理器

2 个文件

SKILL.md

来源说明 · 只读预览

返回搜索返回安装

Tool	Provider	Purpose
`brave_web_search`	brave-search	Find alternative solutions, compare tools
`get_library_docs`	context7	Get documentation for potential replacements
`search_repositories`	github	Find similar OSS implementations
`get_readme`	github	Evaluate alternatives' capabilities

## Feature Inventory

### CLI Commands
- `aoc issues list` — List and filter issues
- `aoc kg build` — Build knowledge graph
- ...

### Internal Modules
- `mcp/` — MCP integration module
- `kg/` — Knowledge graph module
- ...

### Skills/Agents
- `agent-ops-planning` — Multi-iteration planning
- ...

### API Endpoints (if any)
- `GET /api/issues` — List issues
- ...

Query patterns:
- "{feature_name} open source alternatives"
- "best {category} tools 2025 2026"
- "{feature_name} vs {known_alternative}"

#### Alternatives Found
| Tool | Stars/Users | Maintenance | Features | Verdict |
|------|-------------|-------------|----------|---------|
| {name} | {count} | {active/stale} | {comparison} | Better/Same/Worse |

Query patterns:
- "topic:{category} language:{lang}"
- "{feature} implementation"

Verdict	Criteria	Required Evidence
KEEP	Unique value, no better alternative	Documented search showing no alternatives
REMOVE	Better alternatives exist OR unused	Links to superior tools OR zero usage
SIMPLIFY	Overcomplicated for actual use	Usage analysis showing unused options
EXTRACT	Valuable but belongs in separate package	Clear boundaries, standalone value
MERGE	Duplicates another feature	Code overlap analysis
DELEGATE	External tool does it better	Feature comparison table

# Feature Justification Audit Report

**Generated:** {date}
**Project:** {name}
**Scope:** {what was analyzed}

## Executive Summary

| Verdict | Count | % of Features |
|---------|-------|---------------|
| KEEP | N | X% |
| REMOVE | N | X% |
| SIMPLIFY | N | X% |
| EXTRACT | N | X% |
| MERGE | N | X% |
| DELEGATE | N | X% |

**Estimated impact:**
- Lines of code removable: ~N
- Dependencies removable: N
- Maintenance burden reduction: {estimate}

---

## Critical Findings

### 🗑️ Must Remove (Better Alternatives Exist)

#### 1. {Feature Name}
**Verdict:** REMOVE → DELEGATE to {alternative}

**Evidence:**
| Source | Finding |
|--------|---------|
| 🌐 brave-search | {alternative} has 10x stars, active maintenance |
| 📚 context7 | {alternative} docs show feature parity |
| 💭 Analysis | Our implementation: 500 LOC, 3 bugs/month |

**Recommendation:** Replace with {alternative}, delete {files}

---

#### 2. {Feature Name}
...

---

### ✂️ Must Simplify (Overcomplicated)

#### 1. {Feature Name}
**Verdict:** SIMPLIFY

**Evidence:**
| Aspect | Current | Actual Usage |
|--------|---------|--------------|
| Config options | 15 | 3 used |
| Modes | 4 | 1 used |
| LOC | 800 | ~200 needed |

**Recommendation:** Remove unused options, reduce to core functionality

---

### 📦 Extraction Candidates

#### 1. {Feature Name}
**Verdict:** EXTRACT as `{package-name}`

**Rationale:**
- Standalone value outside this project
- Clear API boundary
- Could benefit other projects

---

### ✅ Justified to Keep

#### 1. {Feature Name}
**Verdict:** KEEP

**Justification:**
- Unique functionality: {description}
- Alternatives evaluated: {list}
- Why alternatives don't work: {reasons}
- Usage: {metrics}

---

## Code Duplication Analysis

| Pattern | Locations | Action |
|---------|-----------|--------|
| {description} | file1.py, file2.py | Extract to {module} |

---

## Recommended Action Plan

### Immediate (This Week)
1. Remove {feature} — replace with {alternative}
2. Delete {unused files}

### Short-Term (2 Weeks)
1. Simplify {feature} — remove unused options
2. Extract {module} to shared package

### Medium-Term (1 Month)
1. Evaluate {feature} after usage tracking
2. Complete migration from {old} to {new}

---

## Appendix: Full Feature Analysis

[Detailed per-feature breakdown...]

Note limitation prominently:

⚠️ **Limited Analysis Mode**
MCP tools unavailable. Research based on agent knowledge only.
Findings marked with 💭 require external verification.

Reduce confidence in verdicts:
- KEEP → "KEEP (pending external verification)"
- REMOVE → "REMOVE (verify alternatives exist)"

Recommend follow-up:

## Follow-Up Required

Enable MCP tools and re-run for:
- Alternative verification for {features}
- Documentation comparison for {replacements}

📋 Audit found {N} actionable items:

REMOVE (create CHORE issues):
- Remove {feature} → CHORE "Remove {feature}, replace with {alt}"

SIMPLIFY (create REFAC issues):
- Simplify {feature} → REFAC "Reduce {feature} to core functionality"

EXTRACT (create FEAT issues):
- Extract {module} → FEAT "Extract {module} to standalone package"

Create issues for these? [A]ll / [S]elect / [N]one

Verdict	Issue Type	Discussion Level	Interview Needed?
REMOVE	CHORE	Minimal (clear action)	No
SIMPLIFY	REFAC	Normal (needs design)	Optional
EXTRACT	FEAT	Extensive (architecture)	Yes
MERGE	CHORE	Normal	Optional
DELEGATE	CHORE	Normal	No

## Issue Candidates

### From REMOVE Verdicts

#### CHORE: Remove {feature}, delegate to {alternative}
**Source:** Audit finding #{N}
**Confidence:** High (clear alternative exists)
**Discussion:** minimal
**Files affected:** {list}
**Estimated effort:** S/M/L

---

### From SIMPLIFY Verdicts

#### REFAC: Simplify {feature} to core functionality
**Source:** Audit finding #{N}
**Confidence:** Medium (requires design decisions)
**Discussion:** normal
**Questions to resolve:**
- Which options to keep?
- What's the minimal viable interface?
**Files affected:** {list}

---

### From EXTRACT Verdicts

#### FEAT: Extract {module} as standalone package
**Source:** Audit finding #{N}
**Confidence:** Low (major architectural change)
**Discussion:** extensive 🎤
**Requires interview to determine:**
- [ ] Package name and scope
- [ ] Public API boundaries
- [ ] Versioning strategy
- [ ] Distribution method (PyPI? npm?)
- [ ] Maintenance ownership

📋 Ready to create {N} issues from audit findings:

  🟢 Quick (minimal discussion):     {count} issues
  🟡 Design needed (normal):         {count} issues  
  🔴 Major decision (extensive 🎤):  {count} issues

Issues marked 🎤 will invoke agent-ops-interview for detailed discussion.

Options:
  [A]ll    — Create all issues
  [S]elect — Choose which to create
  [R]eview — Show issue details first
  [N]one   — Skip issue creation

🎤 Issue {ID}: "{title}" requires detailed discussion.

This issue involves architectural decisions that benefit from 
structured interview to ensure requirements are captured.

Start interview now? [Y]es / [L]ater / [S]kip

## Just Completed
- Feature audit report: `.agent/docs/feature-audit-report.md`
- Created {N} issues from audit findings

## Doing Now
- {Interview ID} if extensive discussion in progress

## Up Next
- Review created issues in priority order
- {First high-priority issue from audit}

Check	Tool	Finding
Alternatives	🌐 brave-search	PyYAML (3.2k⭐), ruamel.yaml (1.1k⭐), strictyaml (1.4k⭐)
Docs comparison	📚 context7	ruamel.yaml handles all our syntax + YAML 1.2
Similar projects	🔍 github	No project builds custom YAML parser
Our code	💭 Analysis	450 LOC, 2 bugs last month, 3 call sites

name	prove-your-worth
description	Ruthlessly audit project features for justification. Challenge every feature to prove its value with evidence or face removal. Uses MCP tools for research.
category	analysis
invokes	["agent-ops-tasks","agent-ops-context-map","agent-ops-interview"]
invoked_by	["User request","agent-ops-improvement-discovery"]
state_files	{"read":["focus.md","map.md","constitution.md"],"write":["focus.md"]}
outputs	[".agent/docs/feature-audit-report.md"]

name	prove-your-worth
description	Ruthlessly audit project features for justification. Challenge every feature to prove its value with evidence or face removal. Uses MCP tools for research.
category	analysis
invokes	["agent-ops-tasks","agent-ops-context-map","agent-ops-interview"]
invoked_by	["User request","agent-ops-improvement-discovery"]
state_files	{"read":["focus.md","map.md","constitution.md"],"write":["focus.md"]}
outputs	[".agent/docs/feature-audit-report.md"]

Classification	Meaning	Scrutiny Level
Core	Essential to project identity	Low (but still verify)
Supporting	Enables core features	Medium
Convenience	Nice-to-have	High
Legacy	Historical, unclear purpose	Very High

prove-your-worth

安装方式

检查来源文件

Prove Your Worth — Feature Justification Audit

Philosophy

Prove Your Worth — Feature Justification Audit

Philosophy

When to Use

MCP Tools Required

Procedure

Phase 1: Feature Inventory

Phase 2: Challenge Each Feature

2.1 Alternative Search [🌐 MCP: brave-search]

2.2 Library Documentation [📚 MCP: context7]

2.3 Similar Implementations [🔍 MCP: github]

2.4 Code Analysis [💭 Agent Analysis]

2.5 Usage Analysis [💭 Agent Analysis]

Phase 3: Verdict Assignment

Evidence Standards

For KEEP Verdict (High Bar)

For REMOVE Verdict

Phase 4: Report Generation

Anti-Patterns (What NOT to Do)

MCP Fallback Behavior

Integration with Issue Tracking

Phase 5: Issue Creation & Discussion Planning

5.1 Issue Mapping

5.2 Generate Issue Candidates

5.3 User Confirmation

5.4 Create Issues

5.5 Interview Invocation

5.6 Update Focus

Example Output

Feature: Custom YAML Config Parser

Success Criteria

prove-your-worth

来源信息

安装方式

检查来源文件

同仓库更多 Skills

同仓库更多 Skills

Prove Your Worth — Feature Justification Audit

Philosophy

Prove Your Worth — Feature Justification Audit

Philosophy

When to Use

MCP Tools Required

Procedure

Phase 1: Feature Inventory

Phase 2: Challenge Each Feature

2.1 Alternative Search [🌐 MCP: brave-search]

2.2 Library Documentation [📚 MCP: context7]

2.3 Similar Implementations [🔍 MCP: github]

2.4 Code Analysis [💭 Agent Analysis]

2.5 Usage Analysis [💭 Agent Analysis]

Phase 3: Verdict Assignment

Evidence Standards

For KEEP Verdict (High Bar)

For REMOVE Verdict

Phase 4: Report Generation

Anti-Patterns (What NOT to Do)

MCP Fallback Behavior

Integration with Issue Tracking

Phase 5: Issue Creation & Discussion Planning

5.1 Issue Mapping

5.2 Generate Issue Candidates

5.3 User Confirmation

5.4 Create Issues

5.5 Interview Invocation

5.6 Update Focus

Example Output

Feature: Custom YAML Config Parser

Success Criteria