skills.rest
doable-me
doable-meOFFICIAL
Explore all skills by this author →
GitHub

ai-testing

Coordinates multi-agent exploratory, functional, and regression testing of web, mobile, and API products.

39|14|Updated Mar 14, 2026
Search Tags:#accessibility#multi-agent#test coverage#regression testing#bug reporting#exploratory testing#qa testing
One-click install
npx skills add https://github.com/doable-me/Doable --skill ai-testing-doable-me
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: ai-testing
Source: https://github.com/doable-me/Doable/tree/main/services/api/src/ai/skills/_system/ai-testing
Command: npx skills add https://github.com/doable-me/Doable --skill ai-testing-doable-me

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Manual QA of applications, websites, and APIs is slow and inconsistent, and ad-hoc AI testing often hallucinates bugs or misses edge cases. This Skill provides a rigorous, evidence-driven multi-agent testing workflow that discovers real defects, documents them with reproducible steps, and produces structured QA reports. ## Core Features & Use Cases - Multi-Agent Test Orchestration: Coordinates 11 specialized agents (Test Planner, Exploratory Tester, Functional Validator, Negative Tester, UI/UX Reviewer, Accessibility, Bug Analyst, Regression, Coverage, Evidence, Triage) with defined roles, handoffs, and verification loops. - Structured Bug Reporting: Generates deduplicated bug reports with severity/priority classification, reproducible step sequences, and mandatory screenshot or log evidence for Major-and-above defects. - Coverage and QA Reporting: Produces a final QA report with a per-module coverage map, risk assessment, untested/blocked area lists, and P0/P1 escalation items. - Use Case: Point the Skill at a staging URL of your SaaS dashboard with test credentials; it maps user journeys, runs exploratory and negative tests, files evidenced bug reports, and delivers a triaged QA report. ## Quick Start Ask the AI to test your application at a given URL with provided credentials and produce a full QA report with bug evidence.

Frequently Asked Questions about ai-testing

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

FAQPage Schema
How do I test a web application with AI agents?▼

Provide the application URL, credentials, and any requirements, then the Test Planner agent builds a test plan and coordinates exploratory, functional, and negative testing agents. Results are compiled into structured bug reports and a final QA report with evidence.

What types of testing does multi-agent QA cover?▼

It covers exploratory, functional, negative, boundary, regression, smoke, UI/UX, and accessibility testing against WCAG 2.1 AA. Each testing type is handled by a specialized agent with defined inputs, outputs, and handoff rules.

Can AI testing work on APIs and mobile apps, not just websites?▼

Yes, it supports REST and GraphQL APIs, native iOS and Android apps, desktop applications, admin panels, and SaaS platforms. Mobile and desktop targets require appropriate tool access for interaction and evidence capture.

How are bug severity and priority classified in QA reports?▼

Severity describes how bad a defect is, while priority describes when it should be fixed, and the two are assigned separately using a defined severity matrix. Every Major-or-above bug requires reproducible steps and linked screenshot or log evidence.

What happens when two testing agents report contradictory results?▼

A verification loop is triggered: the Triage agent flags the contradiction, evidence from both sessions is reviewed, and the scenario is retested. Unreproduced findings are filed as observed-once items needing confirmation rather than being silently discarded.

What are the limitations of AI-driven exploratory testing?▼

It cannot assume credentials, staging environments, or documented flows exist, and it must not fabricate evidence when screenshot capture fails. Potential security vulnerabilities are escalated for human review rather than filed as normal bugs.

Related Skills

Explore More Software Engineering Skills→
cameronapakcameronapak
1

bknd-testing

Automate BKND application tests with isolated in-memory environments.

Community
Intermediate
unblockedunblocked
13

headless-validate-bug

Validate bug reports from Jira, Linear, and GitHub Issues.

Official
Advanced
endorlabsendorlabs
2

endor-review

Automate pre-PR security reviews by scanning changed files for vulnerabilities, secrets, licenses, and compliance issues.

Official
Advanced
fgarofalo56fgarofalo56
6

testing

Run unit, integration, and E2E tests with pytest, Jest, Cypress, and Playwright.

Community
Advanced