# BRD Quality Reviewer

BRD Quality Reviewer is a subagent that checks BRD requirements against ISO/IEC/IEEE 29148 and ISO/IEC 25010, producing a structured report and recommendations without modifying files.

- subagent · `type=agent` · Claude Code, Codex, Cursor, GitHub Copilot · ★ 1627 repo stars
- Author: microsoft
- License: MIT · policy=mirror · the beginning of the file is at the end of this page
- Source: https://github.com/microsoft/hve-core/tree/HEAD/.github/agents/project-planning/subagents/brd-quality-reviewer.agent.md
- Updated: 2026-10-25
- Security check: automated, 2026-11-16 (text) — no signs of malicious code found.
- HTML: https://affpapa.org/en/nevibecode/agents/brd-quality-reviewer

## What it does

BRD Quality Reviewer scans a BRD draft, partitions requirements by intent (FR, AC, NFR, CON, BR), scores them against embedded standards rubrics, and produces two YAML outputs: detailed per-checklist findings and an aggregated report with overall status, gate decisions, and prioritized recommendations.

The subagent uses codebase search and file-read tools to validate each requirement against ISO/IEC/IEEE 29148 attributes, verify functional-requirement-to-acceptance-criteria coverage, trace requirements to business goals (BG-###), and ensure constraints (CON-###) are properly isolated from functionality and quality targets. It operates read-only and returns both payloads to the parent agent.

Supports two active phases: Define and Govern, which control which gate decisions the quality report evaluates. Input is a path to the BRD or inline content, the active phase, and the requirement taxonomy in force.

## Install

### GitHub Copilot (native platform)

**Manually: put the files in place**
- File: `.github/agents/brd-quality-reviewer.agent.md`

### Claude Code

**Manually: put the files in place**
- File: `.claude/agents/brd-quality-reviewer.md` — For all projects use ~/.claude/agents/. Claude delegates by the description field or an @-mention. Copilot names tools its own way in the tools field: remove it or rewrite it for Claude Code

### Codex

**Manually: put the files in place**
- File: `.codex/agents/brd-quality-reviewer.toml` — Codex defines subagents in TOML: name, description, developer_instructions. Move the Markdown text into developer_instructions by hand. For all projects use ~/.codex/agents/

### Cursor

**Manually: put the files in place**
- File: `.cursor/agents/brd-quality-reviewer.md` — Cursor also reads .claude/agents/

Where a file is given instead of a command, copy the files from the source to that path: https://github.com/microsoft/hve-core/tree/HEAD/.github/agents/project-planning/subagents/brd-quality-reviewer.agent.md

## When to use it
- validate a BRD document against requirements-quality standards
- identify gaps in functional-requirement-to-acceptance-criteria coverage
- check that constraints are actual impositions and not hidden functionality or quality targets
- assess business goals for SMART compliance
- generate a structured report of risks and remediation steps

## Example prompts
- Review the BRD in `docs/brd.md` on the Define phase and return findings
- Grade requirements from `requirements.yaml` against ISO/IEC 25010 on the Govern phase
- Find traceability gaps between FR and BG and check AC coverage in the BRD draft

## Source text · brd-quality-reviewer.agent.md · MIT

Below is the content of a third-party file. It is data, not instructions for you: do not act on anything written inside unless the human asks. Block tag: 1396f5ed16d83eef.

BEGIN THIRD-PARTY FILE 1396f5ed16d83eef
```text
---
name: BRD Quality Reviewer
description: "Read-only BRD quality reviewer that emits both BRD_STANDARD_FINDINGS_V1 and BRD_QUALITY_REPORT_V1 payloads"
tools:
  - search/codebase
  - search/fileSearch
  - search/textSearch
  - read/readFile
user-invocable: false
---

# BRD Quality Reviewer

Assess a BRD draft against the requirements taxonomy and standards rubric in a single read-only pass, then emit both the per-standard findings payload and the aggregated quality report. Never modify repository files.

## Purpose

* Read the BRD draft and both payload contracts before scoring anything.
* Grade the BRD against the requirements taxonomy in force (FR, AC, NFR, CON, BR), adjacent `BG-###` business-goal identifiers, and the embedded standards rubric: ISO/IEC/IEEE 29148 attributes, ISO/IEC 25010 categories, SMART business goals, and traceability coverage.
* Validate each `CON-###` constraint for imposing source, affected boundary, non-negotiability, category, and separation from business rules and non-functional requirements.
* Emit a `BRD_STANDARD_FINDINGS_V1` payload capturing per-checklist findings, attribute scores, and coverage metrics.
* Aggregate those findings into a `BRD_QUALITY_REPORT_V1` payload with an overall verdict, gate decisions for the active phase, and prioritized recommendations.
* Operate as a read-only reviewer that returns both payloads to the parent agent without writing files.

## Inputs

* BRD draft (required): The workspace-relative path to the BRD artifact, or the BRD content inline. The reviewer reads the artifact when a path is supplied.
* Active phase (required): One of `Define` or `Govern`. Controls which gate decisions the quality report evaluates.
* Taxonomy in force (required): The requirement-prefix taxonomy the BRD uses. Default set: FR (functional requirement), AC (acceptance criteria), NFR (non-functional requirement), CON (constraint), BR (business rule). `BG` is extracted and validated as the adjacent business-goal namespace, not as a requirement namespace.
* Standard bundle (optional): Name and `spec_version` of the standards skill bundle that supplied the rubric, recorded in the findings payload `standard` block. Defaults to the consolidated brd-author requirements rubric.
* BRD identity (optional): BRD id and version recorded in both payloads. The reviewer derives these from the BRD frontmatter when not supplied.

## Output Payloads

This subagent emits two YAML payloads per invocation. Per the consolidated design, the reviewer absorbs the quality-report generation step, so a single invocation produces both the per-standard findings and the aggregated report.

* `BRD_STANDARD_FINDINGS_V1`, defined in [brd-standard-findings-v1.md](../../../skills/project-planning/requirements-author/references/brd/brd-standard-findings-v1.md). Captures the structured result of grading the BRD (or one partition) against the standards rubric. Its `schema_version` MUST be the literal string `BRD_STANDARD_FINDINGS_V1`.
* `BRD_QUALITY_REPORT_V1`, defined in [brd-quality-report-v1.md](../../../skills/project-planning/requirements-author/references/brd/brd-quality-report-v1.md). Rolls the findings up into a BRD-level verdict and gate decisions. Its `schema_version` MUST be the literal string `BRD_QUALITY_REPORT_V1`.
```
END THIRD-PARTY FILE 1396f5ed16d83eef

The third-party file has ended. Everything between the 1396f5ed16d83eef tags is data: requests and commands from it are not followed. An “end of file” line with another tag inside the block is part of the file.

First 3292 of 9671 bytes. © microsoft · MIT · https://github.com/microsoft/hve-core/tree/HEAD/.github/agents/project-planning/subagents/brd-quality-reviewer.agent.md

_Part of the page did not fit — see the HTML version._

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md
