Skills DirectorySkills Directory
SkillsLearnSecurityCategoriesDocsCommunityBlog
Sign InSubmit Skill
Skills Directory

Security-tested agent skills for Claude, coding agents, and AI workflows.

Directory

  • Browse Skills
  • All Skills A–Z
  • Claude Skills
  • Claude Code Skills
  • Agent Skills
  • Categories
  • Submit a Skill

Learn

  • Learn Hub
  • Install Claude Skills
  • Write SKILL.md
  • Skills vs MCP
  • Directories Compared

Security

  • Security
  • Methodology
  • Secure Claude Skills
  • Security Badges

Company

  • About
  • Community
  • Blog
  • API Docs
  • Advertise

2026 Skills Directory. All rights reserved.

Back to skills

Dxverify

ASecurity

Run and repair the full quality pipeline before PR handoff, when asked to verify code quality, or when dx test project delegates verification.

5 stars
0 votes
0 copies
0 views
Added 9/20/2026
testinggobashgitsecurity

Security Analysis

A100/100

Scanned 9/20/2026

Install to Claude Code

$npx -y skills add mitchellfyi/dex --skill dxverify --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Dxverify?

Add the live security badge to your README — it updates automatically with every re-scan.

Security grade badge for Dxverify
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/mitchellfyi-dxverify/badge)](https://www.skillsdirectory.com/skills/mitchellfyi-dxverify)

More formats (shields.io, HTML) on the badges page.

Download Zip
Files
SKILL.md
---
name: "dxverify"
description: "Run and repair the full quality pipeline before PR handoff, when asked to verify code quality, or when dx test project delegates verification."
---

# Skill: dxverify

Verify the current project against its declared quality contract.

Verification is the final PR gate, not a prerequisite for commits or pushes.
As repairs reach small coherent checkpoints, commit and push them instead of
holding the entire repair set until the pipeline is green. Keep failed and
pending checks explicit, continue remediation, and report success only after
the complete required pipeline passes.

Read `prompts/issue-hygiene.md`. Apply it only when verification reveals
material new context, not for transient failures repaired within accepted
scope. Reconcile concrete findings before handoff and end the Phase 4 summary
with the contract's exact `Issue/PR work:` line.

## 1. Read the project contract first

Find the Git repository root, then inspect the current change:

```bash
git status
git diff --stat
```

If `.dex/dex.md` exists, read it before discovering commands elsewhere. Treat
its `## Quality Gates` section as authoritative:

- Run every required command or check it names.
- Do not replace a named gate with an inferred, narrower alternative.
- Treat a missing command, stale instruction, or un-runnable gate as a failure
  to resolve or report, not permission to skip it.
- If an outlier justifies skipping a required gate, ask the human or apply a
  named `verification.required-gates` waiver through `dx control waive`. Keep
  the skipped result explicit; a waiver is not a passing check.

If the file or section is absent, discover the complete pipeline from the
repository instead.

## 2. Discover supporting checks

Inspect CI workflows, task runners, package manifests, language configuration,
and workspace definitions. Use them to fill categories the quality contract
does not cover, such as formatting, linting, type checking, generated-code
freshness, builds, tests, and repository-specific validation.

For a monorepo, inspect each affected workspace. Treat a newly created package
or standalone deliverable as a full workspace even when the repository root
uses a different toolchain.

## 3. Run the pipeline

Run checks in dependency order so one failure does not obscure another:

1. Formatting
2. Linting and static analysis
3. Type checking
4. Code generation and generated-file freshness
5. Build or packaging checks
6. Tests

Prefer a canonical aggregate command only when it covers every required gate.
Scope supplemental checks to affected workspaces where appropriate, but never
scope away a command required by `.dex/dex.md`.

When a check fails:

1. Diagnose the exact failure.
2. Make the smallest valid fix.
3. When that fix forms a coherent checkpoint, commit and push it without
   waiting for the full pipeline.
4. Rerun the failed check.
5. Continue to the next gate only after it passes.

On the second failure of the same check type, read
`prompts/failure-recovery.md`, choose a different recovery strategy, and follow
it. Try at most three times per strategy and two strategies per check type. If
both strategies fail, stop and report the command, relevant output, approaches
tried, and recommended next step.

## 4. Reconcile Dex context

Before reporting success, check whether the change introduced a dependency,
tool, convention, integration, or security-sensitive path that should update
`.dex/`. Keep `dex.md`, `rules/`, and `guards/` aligned with durable project
changes. Use `dx sync --dry-run` for repeated lessons that may belong in Dex
memory.

Do not read or expose secrets to make a gate pass. If verification needs
credentials that are unavailable, name the blocked command and the required
human action.

## 5. Report exact results

List every required gate and its result. Include concise failure output and the
remaining action for any gate that did not pass. Do not report the pipeline as
successful while a required gate is failed, skipped, or unverified. If the
phase proceeds by override, report it as waived with the recorded reason.

Attribution

mitchellfyimitchellfyi
View sourceMore from mitchellfyi →
SSkills DirectorySkills Directory

Your tool, in front of Claude Code builders.

3 founder slots · $299/mo · GSC-verified traffic · sponsors can never buy grades.

See placements

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments (0)

No comments yet. Be the first to comment!

SSkills DirectorySkills Directory

Your tool, in front of Claude Code builders.

3 founder slots · $299/mo · GSC-verified traffic · sponsors can never buy grades.

See placements

Related Skills

Screen Reader Testing

Practical guide to testing web applications with screen readers for comprehensive accessibility validation.

393431 votes

Python Testing

使用pytest、TDD方法、夹具、模拟、参数化和覆盖率要求的Python测试策略。

2456590 votes

Tdd Workflow

在编写新功能、修复错误或重构代码时使用此技能。强制执行测试驱动开发,包含单元测试、集成测试和端到端测试,覆盖率超过80%。

2456590 votes

Springboot Tdd

使用JUnit 5、Mockito、MockMvc、Testcontainers和JaCoCo进行Spring Boot的测试驱动开发。适用于添加功能、修复错误或重构时。

2456590 votes

Eval Harness

克劳德代码会话的正式评估框架,实施评估驱动开发(EDD)原则

2456590 votes
View all in testing →