Token导航 LogoToken导航TokenDH.com
前端设计只读github未标认证来源可访问许可证需确认审计通过

quality-enforcement质量执行

Agent Skill

quality-enforcement 用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要围绕仓库状态、代码变更或协作事项进行整理时使用。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。

总安装

282

周安装

12

GitHub Stars

24

下载量

99
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

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:quality-enforcement(质量执行)
来源仓库:https://github.com/noobygains/godmode
仓库路径:skills/quality-enforcement
安装命令:
npx skills add https://github.com/noobygains/godmode --skill quality-enforcement
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/noobygains/godmode --skill quality-enforcement

简介

quality-enforcement 用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息,适合在开发流程中整合项目状态与变更细节。

  • 适用于需要围绕仓库活动、代码提交或协作事项进行整理的场景。
  • 通过标准安装方式接入后,可调用其接口获取相关元数据或工单信息。
  • 使用前请确认访问权限、维护状态及是否允许执行网络或文件操作。
  • 适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。

SKILL.md

Quality Enforcement

Overview

Quality enforcement is automated, not aspirational. If it can be checked by a machine, it must be.

Core principle: A quality check that is not enforced in CI does not exist.

No exceptions. No workarounds. No shortcuts.

The Prime Directive

NO CODE LANDS WITHOUT ALL QUALITY CHECKS PASSING

If a check fails, fix the code. Never disable the check. Never bypass CI.

When to Use

Always before:

  • Committing code
  • Opening a pull request
  • Merging to trunk
  • Releasing to production

Especially when:

  • "Just suppress the linter this once" (never)
  • "CI is too slow, merge manually" (fix CI, do not skip it)
  • "Type errors but it works at runtime" (fix the types)
  • "Coverage dropped but the critical paths are tested" (restore coverage)

The Entry Protocol

BEFORE any PR or merge:

1. LINT: Zero errors, zero warnings
2. TYPES: Zero type errors (strict mode)
3. TESTS: All passing, coverage meets threshold
4. BUILD: Clean build, no warnings
5. SIZE: Bundle/binary within budget (if applicable)
6. DEPS: No known vulnerabilities (critical/high)
7. COMPLEXITY: No functions exceeding complexity threshold

Any check fails = code is not ready. Fix before advancing.

Quality Check Reference

Check 1: Linting

Standard: Zero errors AND zero warnings
SettingValueRationale
Errors0Non-negotiable
Warnings0Warnings become errors you learn to ignore
Config committedYesConsistent across all contributors
CI enforcedYesLocal overrides are irrelevant

Warnings are tomorrow's errors. Either fix them or adjust the rule. Never tolerate warnings.

Disable a rule? Only if the team explicitly agrees the rule is inappropriate for this project. Document the rationale in the config file. Never disable inline for convenience.

Check 2: Type Safety

Standard: Zero type errors, strict mode enabled

TypeScript:

{
  "compilerOptions": {
    "strict": true,
    "noUncheckedIndexedAccess": true,
    "noImplicitReturns": true,
    "noFallthroughCasesInSwitch": true,
    "forceConsistentCasingInFileNames": true
  }
}

Python: mypy or pyright with strict mode.

Anti-PatternProblemFix
any typeDisables type checkingUse specific types or generics
// @ts-ignoreConceals real errorsFix the type error
# type: ignoreSame problemFix the type error
Non-strict modeFalse sense of safetyEnable strict from day one

any is a type error you chose not to resolve. Every any weakens the type system for everything it touches.

Check 3: Test Coverage

Standard: Coverage threshold that never decreases
MetricFloorTarget
Line coverage80%90%+
Branch coverage70%85%+
New code coverage90%100% (aspire to)

Coverage thresholds are a ratchet. They increase, never decrease. Configure CI to fail if coverage drops below the current level.

Coverage is necessary but not sufficient. 100% coverage with poor assertions is worse than 70% coverage with rigorous assertions. Coverage checks combine with TDD discipline (godmode:test-first).

Check 4: Build

Standard: Clean build, zero warnings
  • Build must complete successfully
  • Zero compiler/build warnings
  • Output matches expected structure
  • No missing dependencies at build time

Check 5: Bundle Size Budget (Frontend)

Standard: Total bundle size within defined budget
TargetBudgetTool
Initial JS (compressed)<200KBwebpack-bundle-analyzer
CSS<50KBPurgeCSS check
ImagesOptimized, WebP/AVIFimagemin
Total page weight<1MBLighthouse

Set the budget. Enforce in CI. When budget is exceeded, analyze the bundle contents and eliminate or split.

Check 6: Dependency Audit

Standard: Zero critical or high vulnerabilities
LanguageCommandCI Integration
JavaScriptnpm audit --audit-level=highFail on high+
Pythonpip-audit or safety checkFail on high+
Gogovulncheck./...Fail on any
Rustcargo auditFail on any

Also verify:

  • No unnecessary dependencies (is every dep actually used?)
  • No duplicate dependencies (different versions of same package)
  • Dependencies are maintained (last update within 12 months)

Check 7: Complexity Metrics

Standard: No function exceeds complexity threshold
MetricThresholdTool
Cyclomatic complexity<10 per functionESLint (complexity rule), radon, gocyclo
Function length<50 linesLinter rules
File length<400 linesLinter rules
Parameters<5 per functionLinter rules

When complexity exceeds threshold: Refactor. Extract functions. Simplify conditionals. Never raise the threshold.

Check 8: Dead Code

Standard: No unused exports, variables, or dependencies
WhatTool
Unused exportsts-prune, knip
Unused dependenciesdepcheck, knip
Unused variablesLinter (no-unused-vars)
Unreachable codeLinter, type checker

Dead code is misleading code. It implies something depends on it. Remove it.

CI Pipeline Template

# Minimum quality enforcement pipeline
quality-checks:
  steps:
    - name: Lint
      run: npm run lint
    - name: Type Check
      run: npm run typecheck
    - name: Test
      run: npm test -- --coverage
    - name: Coverage Check
      run: check-coverage --threshold 80
    - name: Build
      run: npm run build
    - name: Bundle Size
      run: bundlesize
    - name: Audit
      run: npm audit --audit-level=high

All checks run on every PR. All must pass before merge.

Cognitive Traps

RationalizationTruth
"Just a lint warning, not an error"Warnings you ignore become errors you miss.
"Type error but it works at runtime"Types prevent the runtime error you have not encountered yet.
"Coverage dropped 1%, not a big deal"1% per PR = 50% in a year. Ratchets do not go down.
"Skip CI, I tested locally"Local environments differ from CI. That is why CI exists.
"Bundle grew because we added features"Features should replace or split, not only add.
"Vulnerability is in a dev dependency"Dev deps run in CI. CI has secrets. Still a risk.
"Function is complex but readable"Complexity limits exist because readability is subjective.
"Dead code might be needed later"Git remembers. Delete it. Restore from history if needed.

Guardrails - HALT and Fix

  • Disabling lint rules inline without documented rationale
  • @ts-ignore or # type: ignore without an accompanying issue
  • Coverage threshold lowered in config
  • CI skipped or overridden for merge
  • any types spreading through codebase
  • Bundle size growing without investigation
  • Warnings treated as acceptable
  • Audit failures dismissed because "it's a dev dependency"

All of these mean: The check is broken. Fix the check before writing more code.

Integration

Complements:

  • godmode:test-first — Tests are one check among many
  • godmode:completion-gate — Quality checks are verification evidence
  • godmode:project-bootstrap — Checks configured at project setup
  • godmode:security-protocol — Dependency audit is a security check
  • godmode:performance-tuning — Bundle size is a performance check

The Bottom Line

Every quality check automated in CI. Every check passing before merge. No exceptions.

If a check can be bypassed, it will be bypassed. Make it impossible to bypass.

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

36.73%
按下载量换算36

Claude

27.33%
按下载量换算27

Cursor

16.78%
按下载量换算17

Gemini CLI

10.16%
按下载量换算10

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。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills