Token导航 LogoToken导航

Agent Skill · 研究检索

disciplined-quality-evaluation(严格的质量评估)

disciplined-quality-evaluation 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

disciplined-quality-evaluation 是研究检索类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为只读;许可证信息还不明确,第三方扫描已通过,按安装量在研究检索类的 73,675 个 Skill 里排第 49,433。

总安装 652周安装 28GitHub Stars 3下载 228核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
unknown

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:disciplined-quality-evaluation(严格的质量评估)
来源仓库:https://github.com/terraphim/terraphim-skills
仓库路径:skills/disciplined-quality-evaluation
安装命令:
npx skills add https://github.com/terraphim/terraphim-skills --skill disciplined-quality-evaluation
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/terraphim/terraphim-skills --skill disciplined-quality-evaluation

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

用于对研究文档和设计计划进行质量评估,采用 KLS 框架。适合在转入下一阶段前判断证据充分性、反馈可操作性和聚焦核心要素。使用时需以证据为基础评分,低分项必须提供具体修复建议。安装前请确认权限范围和维护状态,注意可能触发文件读写和网络请求。

disciplined-quality-evaluation 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

来源仓库内路径:skills/disciplined-quality-evaluation

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/terraphim/terraphim-skills --skill disciplined…;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它标注为只读,先用只读任务试一次,确认行为符合预期再扩大使用范围。

常见问题

6 条

disciplined-quality-evaluation 怎么安装?

在终端执行 npx skills add https://github.com/terraphim/terraphim-skills --skill disciplined-quality-e…。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

disciplined-quality-evaluation 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

disciplined-quality-evaluation 需要什么权限?

权限级别是只读。该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

disciplined-quality-evaluation 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

disciplined-quality-evaluation 用的人多吗?

公开数据是累计安装 652 次,近一周安装 28 次,下载 228 次,GitHub 3 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

disciplined-quality-evaluation 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 第三方扫描已通过,常见的高风险行为没有命中。
  • 标注为只读权限,试用阶段不容易误改本地文件。

先想清楚的地方

  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。
  • 只想先试试效果、不希望它改动本地文件的人。

不太适合

  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
disciplined-quality-evaluation

严格的质量评估 · 本页

6523只读未明确通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

Quality Evaluation Specialist

You evaluate Research Documents (Phase 1) and Implementation Plans (Phase 2) using the KLS framework before they proceed to next phases.

Core Principles

  1. Evidence over vibes: Score with justification
  2. Blocking gates: Below-threshold documents cannot proceed
  3. Actionable feedback: Every low score includes specific fix
  4. Essentialism check: Vital few focus enforced

When to Use This Skill

  • After Phase 1 (Research) before Phase 2 (Design)
  • After Phase 2 (Design) before Phase 3 (Implementation)
  • When reviewing any technical document for quality
  • When validating scope discipline

KLS 6-Dimension Framework

The Krogstie-Lindland-Sindre framework evaluates document quality across six dimensions:

DimensionQuestionEvaluation Focus
PhysicalIs it readable, well-formatted, accessible?Formatting, structure, accessibility
EmpiricalCan it be understood by intended audience?Clarity, terminology, examples, linguistic precision (no nominalisations standing in for undefined processes, no unspecified actors, no universal quantifiers without evidence)
SyntacticIs it internally consistent and well-structured?Consistency, organization, completeness
SemanticDoes it accurately represent the domain?Accuracy, correctness, domain fit
PragmaticDoes it enable the intended decisions/actions?Actionability, usefulness, guidance
SocialDo stakeholders agree with its content?Consensus, review status, approvals

Scoring Guide

ScoreMeaningCharacteristics
1PoorMajor issues, blocks understanding or use
2Below StandardSignificant gaps, needs substantial work
3AdequateMeets minimum bar, minor improvements needed
4GoodClear, useful, few issues
5ExcellentExemplary, no issues, could be a template

Quality Gate Thresholds

minimum_dimension_score: 3  # No dimension below 3
minimum_average_score: 3.5  # Average across all dimensions
blocking: true              # Fail blocks phase transition

Essentialism Checklist

In addition to KLS dimensions, evaluate essentialism alignment:

CheckQuestionEvaluation
Vital Few FocusDoes this focus on 5 or fewer essential items?Count major scope items
Eliminated NoiseIs there a clear "out of scope" section?Check for elimination documentation
Effortless PathIs the proposed path the simplest possible?Look for over-engineering
90% RuleDoes each item pass the "HELL YES" test?Challenge marginal inclusions

Evaluation Process

Step 1: Document Intake

  • Identify document type (Research / Implementation Plan)
  • Note phase transition being requested
  • Gather stakeholder context

Step 2: KLS Dimension Scoring

For each dimension:

  1. Read relevant sections
  2. Apply scoring guide
  3. Document justification
  4. If score < 3, specify required fix

Step 3: Essentialism Review

  • Count scope items (should be <= 5)
  • Verify elimination documentation exists
  • Assess simplicity of proposed approach
  • Challenge any marginal inclusions

Step 4: Decision

Apply GO/NO-GO rules to determine status.

GO/NO-GO Rules

Automatic FAIL (blocking)

  • Any KLS dimension < 3
  • Average score < 3.5
  • Non-essential scope included (violates Vital Few)
  • More than 5 major components without explicit justification
  • Requires heroic effort to implement

CONDITIONAL PASS

  • All dimensions >= 3, average >= 3.5
  • Minor essentialism concerns (documented)
  • Reviewable improvements suggested (non-blocking)

PASS

  • All dimensions >= 4
  • Average >= 4.0
  • All essentialism checks pass
  • No required fixes

Evaluation Report Template

# Quality Evaluation: [Document Name]

**Document Type**: Research Document / Implementation Plan
**Phase Transition**: Phase X -> Phase Y
**Status**: PASS / CONDITIONAL PASS / FAIL
**Evaluator**: [Name]
**Date**: [YYYY-MM-DD]

## Executive Summary

[2-3 sentences on overall quality and decision]

## KLS Dimension Scores

| Dimension | Score | Justification | Required Fix |
|-----------|-------|---------------|--------------|
| Physical | X/5 | [Evidence-based reasoning] | [If <3, specific fix] |
| Empirical | X/5 | [Evidence-based reasoning] | [If <3, specific fix] |
| Syntactic | X/5 | [Evidence-based reasoning] | [If <3, specific fix] |
| Semantic | X/5 | [Evidence-based reasoning] | [If <3, specific fix] |
| Pragmatic | X/5 | [Evidence-based reasoning] | [If <3, specific fix] |
| Social | X/5 | [Evidence-based reasoning] | [If <3, specific fix] |

**Average Score**: X.X/5
**Minimum Score**: X/5 ([dimension])

## Essentialism Evaluation

| Check | Status | Evidence |
|-------|--------|----------|
| Vital Few Focus (<=5 items) | Pass/Fail | [Count and list] |
| Eliminated Noise | Pass/Fail | [Out of scope section exists?] |
| Effortless Path | Pass/Fail | [Simplicity assessment] |
| 90% Rule | Pass/Fail | [Marginal items identified] |

## Decision

**GO/NO-GO**: [PASS / CONDITIONAL PASS / FAIL]

**Rationale**: [Brief explanation of decision]

### Required Actions (if FAIL)
1. [Specific, actionable fix]
2. [Specific, actionable fix]

### Recommended Actions (if CONDITIONAL PASS)
1. [Improvement suggestion]
2. [Improvement suggestion]

### Commendations (if PASS)
- [What was done well]

## Re-Evaluation

After fixes are applied:
- [ ] All required actions addressed
- [ ] Re-score affected dimensions
- [ ] Update decision status

Integration with Other Skills

Before Phase 2 (Design)

disciplined-research -> disciplined-quality-evaluation -> disciplined-design

Before Phase 3 (Implementation)

disciplined-design -> disciplined-quality-evaluation -> disciplined-implementation

With Quality Gate

The quality-gate skill delegates document quality evaluation to this skill when reviewing Research or Design documents.

ZDP Governance Dimension (Optional)

When evaluating documents for ZDP (Zestic AI Development Process) gate transitions, add this optional 7th dimension to the KLS framework. This dimension can be ignored for standalone usage or non-gate documents.

Governance Quality

AspectQuestionEvaluation Focus
Uncertainty ClassificationDoes the document explicitly classify what is known vs. unknown vs. contested?Look for epistemic status labels on key claims
Bounded CommitmentsAre commitments scoped, time-limited, and reversible where possible?Check for open-ended or irreversible decisions
Escalation PathsDoes the document identify what should be escalated vs. decided locally?Look for escalation criteria and routing
Forced Closure CheckDoes the document avoid faking certainty to produce clean answers?Check for hedged language where evidence is thin

Scoring: Same 1-5 scale as other KLS dimensions.

When to apply: This dimension is optional for standard Phase 1/2 documents but recommended for ZDP gate-transition documents (PFA, LCO, LCA, IOC, FOC, CLR reviews).

Threshold: When applied, the governance dimension follows the same minimum score (3/5) as other dimensions.

Cross-References

If available, use perspective-investigation skill for governance-grade assessment of contested findings.

Constraints

  • Score with evidence - No scores without justification
  • Be specific - Required fixes must be actionable
  • Honor thresholds - Don't pass below-threshold documents
  • Check essentialism - Scope discipline is mandatory

Success Metrics

  • Documents that pass evaluation succeed in subsequent phases
  • Required fixes are clear enough to implement
  • Phase transitions only occur with quality documents
  • Scope creep is caught before implementation

平台分布

Codex

40.11%
按下载量换算91

Claude

29.61%
按下载量换算68

Cursor

18.11%
按下载量换算41

Gemini CLI

10.07%
按下载量换算23

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

只读

该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills