skills.rest
threadedstack
threadedstackOFFICIAL
Explore all skills by this author →
WebsiteGitHub

verify-completion

Validates task completion by running type checks, tests, and deliverable audits before reporting done.

Updated Jan 4, 2026
Search Tags:#code quality#testing#verification#type checking#git diff#ci#completion audit
One-click install
npx skills add https://github.com/threadedstack/threadedstack --skill verify-completion-threadedstack
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: verify-completion
Source: https://github.com/threadedstack/threadedstack/tree/main/.claude/skills/verify-completion
Command: npx skills add https://github.com/threadedstack/threadedstack --skill verify-completion-threadedstack

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? AI coding agents often claim work is finished without evidence, leaving broken types, failing tests, and half-implemented features. This Skill enforces a mandatory, evidence-based verification workflow before any task can be reported as complete. ## Core Features & Use Cases - Deliverable Enumeration: Lists every requested item with explicit DONE/INCOMPLETE/NOT STARTED status so nothing is silently dropped. - Evidence-Based Verification: Runs type checks, unit tests, and integration tests per affected repo, requiring actual command output as proof rather than summaries. - Anti-Laziness Scanning: Greps your own diff for TODO, FIXME, placeholder, and stub patterns, plus audits adjacent code for broken imports or missing error handling. - Use Case: After implementing a multi-repo feature touching backend and frontend, run this verification to produce a structured completion report proving types pass, tests are green, and no deferred work remains. ## Quick Start Ask the agent to run the completion verification before reporting the current implementation task as done.

Frequently Asked Questions about verify-completion

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I verify an AI coding agent actually finished the task?▼

Require evidence-based verification: enumerate every requested deliverable with a status, run type checks and tests per affected repo, and paste actual command output. Never accept a summary claim of success without the raw results.

How to check for unfinished TODO or placeholder code in a git diff?▼

Run git diff piped to grep with a pattern matching TODO, FIXME, HACK, placeholder, stub, and similar markers, filtering to added lines. Any match in your own changes means the work must be completed before claiming done.

When should integration tests run before marking work complete?▼

Run integration tests whenever changes touch backend API endpoints, proxy routing, database schemas, authentication, or cross-repo behavior. If you skip them, you must explicitly state why they are not applicable.

Can this verification workflow commit or push changes to git?▼

No. The workflow strictly uses read-only git commands like git diff and git status to identify changed files. It never commits, pushes, or modifies git history during verification.

What happens if type checks or tests fail during verification?▼

The work is not done. You must fix the type errors or failing tests, re-run the checks, and repeat until clean. Skipping a failing step or claiming success without passing output is explicitly prohibited.

Related Skills

Explore More Software Engineering Skills→
MinhDuyDEVMinhDuyDEV

spec-driven-development

Transform ambiguous change requests into structured, testable specifications with acceptance criteria.

Community
Advanced
kenhiakenhia

land

Commit, push, create a pull request, and generate a handoff summary.

Community
Advanced
AndyHsuTWAndyHsuTW

django-verification

Automate Django project verification with linters, security scanners, and tests.

Community
Advanced
YusufkotavomYusufkotavom

testing-quality-gates

Run lint, type-check, and tests across scoped changes for pre-commit or CI workflows.

Community
Advanced