verification

$npx mdskill add JasonxzWen/harness-hub/verification

Collects evidence from repository gates to verify completion claims.

  • Solves verifying that a change passes all required repository gates.
  • Depends on repository manifests, CI configs, and existing scripts.
  • Discovers actual gates from project instructions and classifies by task boundary.
  • Records exact commands, exit status, scope, and freshness for each gate.

SKILL.md

.github/skills/verificationView on GitHub ↗
---
name: verification
description: Load when claiming completion requires the repository's actual deterministic build, type, lint, test, diff, smoke, and artifact gates; use diagnose while a failure's root cause is unknown.
license: MIT
---

# Verification

Verification is evidence collection, not a generic command recipe or a verdict generator.

## Discover The Gates

Discover the project's actual gates from repository instructions, package/task manifests, CI configuration, existing scripts, and nearby tests. Do not invent commands, coverage thresholds, or required tools. Prefer the same entry points maintainers and CI already use.

Classify checks by the accepted task boundary:

- targeted behavior tests for the changed surface
- nearest relevant suite
- build, type, lint, format, schema, or generated-artifact checks
- broader repository validation when blast radius warrants it
- smoke or manual checks that prove the real user path
- Git diff/status checks for unintended files and whitespace errors

Run independent read-only checks in parallel only when they do not compete for mutable artifacts.

## Execute And Record

For every applicable gate, record:

- exact command or reproducible manual procedure
- exit status and concise result
- scope covered
- whether evidence is fresh for the current checkout

Report a required gate that cannot run as `unavailable or unknown`, with the missing dependency or access. Do not translate absence of evidence into a pass.

Deterministic failures cannot be waived by Agent judgment. If the root cause is unclear, use `diagnose`; if a confirmed code behavior needs a test-led fix, use `tdd`. Re-run affected gates after any mutation.

## Artifact And Diff Checks

Inspect produced artifacts when correctness depends on rendering, packaging, generated files, migrations, or browser behavior. A successful command is not evidence that the artifact itself is usable.

Before completion, review current Git status and diff for:

- unintended or missing changes
- stale generated output
- accidental credentials or project-specific data
- incomplete renames and stale references
- whitespace errors

## Handoff

Return the smallest truthful verification table:

| Gate | Evidence | Result |
|---|---|---|
| Project-defined check | Exact command/procedure and scope | pass / fail / unavailable or unknown |

State residual risk and the highest-signal next action. Do not commit, push, publish, merge, or modify remote state merely because verification passed.

More from JasonxzWen/harness-hub

SkillDescription
agent-interaction-auditLoad when the user asks for private agent interaction audits from local Codex/Claude Code traces, task profiles, tool-call decisions, collaboration bottlenecks, prompt/rule or automation-log audits, closeout retrospectives, or evidence-backed Skill/project-rule/Eval/SOP/OKF improvements.
agent-reachLoad when a task needs read-only internet research or content retrieval through an already installed Agent Reach CLI; diagnose available backends first, never install, upgrade, authenticate, configure, or copy credentials on the user's behalf.
animation-vocabularyLoad when a user describes a Web animation or motion effect without knowing its name and needs the precise term or close alternatives; use only for naming, not design, implementation, or review.
apple-designLoad when a frontend task needs Apple-platform design guidance or physical, gesture-driven, velocity-aware, interruptible interaction; skip generic visual styling and use review-animations for report-only motion review.
beautiful-feishu-whiteboardLoad when creating or restyling a polished editable Feishu or Lark whiteboard, infographic, diagram, poster, or visual explainer; do not load for generic SVGs, static images, or other whiteboard platforms.
clone-websiteLoad when the user explicitly asks to clone, reverse-engineer, or rebuild a website they own or are authorized to reproduce; do not load for generic redesign, inspiration, scraping, phishing, impersonation, or copying third-party brands.
codebase-to-courseLoad when a user wants a source-backed interactive HTML course that explains how a codebase works; do not load for chat tutoring, repository review, articles, slide decks, or general browser artifacts.
context-guidanceFixture skill used to represent an additional standard skill context surface.
decision-uiLoad when Codex 即将要求用户在 2-3 个互斥方案中做真实决定、准备询问阻塞性澄清问题或需要确认高影响取舍;将选择转换为宿主原生可点击 UI,并在工具不可用时诚实文本回退。不要用于低风险实现细节、只有一个合理答案的问题、普通信息采集或不需要用户决策的完成汇报。
design-taste-frontendLoad when a task needs anti-template frontend visual direction, design-read calibration, or pre-flight critique for landing pages, portfolios, marketing pages, or redesigns; do not load for dashboards, data tables, multi-step product UI, routine frontend logic, HTML reports, slide decks, or generic code explanation.