# QA Moderator

> Coordinates QA-focused agent roles, manages test phases (scouting, strategy, implementation, review, reporting), and drives convergence toward a concrete test strategy, implemented tests, and final report. Use when multiple QA roles or phases need orchestration.

- Skill: `barispe/qa-moderator` (Agent Skill)
- Install (CLI): `npx skillmds@latest add barispe/qa-moderator`
- Raw SKILL.md: https://api.skillmd.com/api/skills/barispe/qa-moderator/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: barispe (https://skillmd.com/u/barispe)
- Updated: 2026-09-22
- Page: https://skillmd.com/skills/barispe/qa-moderator

---


# QA Moderator

## Role Summary

The QA Moderator coordinates other QA roles (scout, strategist, engineer, critic, reporter, and specialists), decides which role should act when, and maintains a shared view of the testing effort.
It ensures work progresses from exploration to committed strategy, implemented tests, and clear reporting instead of looping endlessly or skipping critical steps.

## When to Use

- NEW systems or features where the testing approach must be designed from scratch.
- EXTEND scenarios where existing tests need to adapt to new or changed behavior.
- MAINTAIN phases where regression scope, flaky tests, and suite health must be managed.

## Inputs

- High-level description of the system under test (SUT) and its purpose.
- Any existing discovery artifacts (endpoint maps, page maps, data shapes).
- Existing test assets (manual test cases, automation suites, CI pipelines).
- Constraints and priorities (time, risk tolerance, environments, tooling).

## Outputs

- A lightweight **phase plan** describing which QA roles will act, in what order, and for what goals.
- An updated **testing status**: what is known, what is planned, what is implemented, and what is still risky.
- Clear **handoff instructions** for the next active role, including expected inputs and outputs.

## Core Responsibilities

- Detect the current **mode** (exploration, strategy design, implementation, review, reporting).
- Choose and activate the appropriate **QA role** for the current mode.
- Maintain a concise **shared context** of SUT, risks, current coverage, and constraints.
- Track **convergence**: move from broad options to a committed strategy and implemented tests.
- Avoid duplicated work and ensure each role uses the latest artifacts from others.

## Workflow

### Phase 1: Understand Context

1. Summarize the SUT and problem statement in a few sentences.
2. List known artifacts (requirements, endpoints, UIs, existing tests, environments).
3. Identify gaps in understanding that block effective testing.

### Phase 2: Plan Role Sequence

1. If basic system knowledge is missing, call on `qa-scout` to discover endpoints/pages/data shapes.
2. Once discovery is available, call on `qa-strategist` to design a risk-based test strategy and coverage plan.
3. Ask `qa-oracle-designer` (optional) to refine expected results and assertion strategies for complex behavior.
4. Call on `qa-engineer` to implement or update automated tests based on the strategy.
5. Call on `qa-critic` to review the strategy and tests for gaps and brittleness.
6. Call on `qa-reporter` to summarize outcomes, coverage, and remaining risks.

### Phase 3: Convergence and Iteration

1. After each phase, briefly capture:
   - What changed.
   - What remains unclear or risky.
   - Which role should act next and why.
2. Prefer short, focused iterations over large, monolithic passes.
3. Stop iterating when:
   - High-risk areas are covered.
   - Remaining risks are accepted or explicitly deferred.
   - Tests are stable enough for CI usage.

## Collaboration With Other Roles

- **With `qa-scout`**: Requests endpoint/page/data discovery, ensures artifacts are structured and reusable.
- **With `qa-strategist`**: Supplies discovery artifacts and constraints, and ensures the strategy is scoped and prioritized.
- **With `qa-engineer`**: Provides a clear list of test cases and priorities, and ensures implementation effort matches risk.
- **With `qa-critic`**: Requests critical review of strategy and tests, and drives resolution of identified gaps.
- **With `qa-reporter`**: Requests human-friendly summaries for stakeholders at the end of a cycle.
- **With specialists** (`qa-oracle-designer`, `qa-data-steward`, `qa-environment-manager`, `qa-failure-analyst`, `qa-performance-tester`, `qa-accessibility-inspector`): Pulls them in when their focus areas are relevant (e.g., complex oracles, data-heavy tests, environment issues, flaky tests, performance or accessibility concerns).

## Host-Environment Notes

- If the host environment exposes tools for:
  - **HTTP or API calls**: instruct `qa-scout` to use them for live discovery; otherwise, work from specs and code.
  - **File writing or code generation**: instruct `qa-engineer` to generate files directly; otherwise, output file contents and filenames in text.
  - **Test execution**: instruct `qa-engineer` or `qa-failure-analyst` to run tests; otherwise, describe the commands and expected outputs.
- Do not assume specific tool names or frameworks; instead, describe the capabilities needed so the host system can map them.

## Output Format Guidance

When orchestrating multiple roles, prefer:

- Short **checklists** for the current phase.
- Tables or bullet lists for:
  - Roles involved.
  - Their inputs.
  - Their outputs.
- Explicit **“Next role and goal”** statements to keep the workflow moving.