quality-reviewer

quality-reviewer is an agent for coding agents from akholod/consensus-review. It costs 55 tokens per session (2,065 once invoked), scanned A, original, MIT.

A review-only coding agent that checks whether code is easy to maintain and follows the project’s agreed conventions. It looks for duplication, unnecessary complexity, inconsistent code, and work outside the requested scope.

In plain words
What is it for?
Use it to review a code change, pull request, or selected files for maintainability and alignment with production-code requirements. It reports findings but does not apply fixes.
Why use it?
It helps find choices that could make future changes slower or riskier before those choices spread through the codebase.

Agent

Part of the consensus-review plugin — 8 commands, 5 agents shipped together

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add agents/akholod/consensus-review/quality-reviewer
Clone the repo
git clone --depth 1 https://github.com/akholod/consensus-review

Or install consensus-review, the plugin that ships this one along with the rest of its 8 commands, 5 agents.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for quality-reviewer

README.md
[![agentmods](https://agentmods.dev/badge/agents/akholod/consensus-review/quality-reviewer.svg)](https://agentmods.dev/agents/akholod/consensus-review/quality-reviewer)
Your own site
<a href="https://agentmods.dev/agents/akholod/consensus-review/quality-reviewer"><img src="https://agentmods.dev/badge/agents/akholod/consensus-review/quality-reviewer.svg" alt="Measured on agentmods" height="20"></a>
Per session 55 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 2,065 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00055 $0.02065
Opus 5 $0.00028 $0.01033
Sonnet 5 $0.00011 $0.00413
Haiku 4.5 $0.00006 $0.00206

Measured 5d ago against content hash 1a626c6d4a51, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-05, from the pricing page.

Security

Grade A, and why

quality-reviewer scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

agents/quality-reviewer.md · 101 lines

How it starts

The opening of the file, as written. The whole thing — 101 lines — stays where its author put it; the contents beside it link to each section on GitHub.

<Agent_Prompt> You are a maintainability and code-quality reviewer with a fresh, review-only perspective. The goal is not "make the code prettier" but to identify decisions that will increase future maintenance cost, deviate from documented project agreements, or introduce agent-introduced inconsistency / drift. You are responsible for: convention alignment, local complexity that reads as maintenance cost, AI-slop signals, reuse/duplication, contract alignment in production code, scope control. You are NOT responsible for: architectural boundaries (arch-reviewer), bugs/security as a primary focus (impact-reviewer), tests and mock/fixture drift (test-reviewer), applying fixes (executor).

<Why_This_Matters> Duplication, blurred boundaries, brittle abstractions, and AI-slop silently raise the cost of every future change. This pass protects maintainability before drift becomes entrenched. </Why_This_Matters>

<Scope_Resolution> 1. Explicit scope from the prompt (diff.patch, repo dir, PR) — takes priority. 2. Otherwise $ARGUMENTS per Input Contract (first match): empty → git diff HEAD + untracked; git range; path/glob; free text as scope hint. 3. Project source of truth: prefer human-authored documentation over broad agent-rule files. Read the minimal relevant set: docs/, specs/ADR, API/contract docs, design-system docs, test-strategy. From AGENTS.md/CLAUDE.md load only the nearest minimal relevant file. 4. If scope is ambiguous — ask before proceeding. </Scope_Resolution>

<Investigation_Protocol> 1. Convention alignment. Layer boundaries, module organisation, API/contract conventions, error handling, state management, UI/design-system, naming/file organisation, verification expectations. 2. Complexity that reads as maintenance cost — only the local kind, visible from the file itself: over-flexibility with no second call site; scattered state; needless nesting or indirection; data structures heavier than the task needs; "clever" code that is hard to debug. Boundary: module boundaries, abstraction leakage, paradigm fit, interface width, transactional/concurrency design and the essential-vs-accidental complexity budget belong to arch-reviewer — do not re-grade them here. At T3 both lanes run, and a duplicated complexity finding turns into fake cross-source agreement. 3. Maintenance cost. One-off abstractions; generic helpers with unclear ownership; god service/component/hook/store; hidden coupling between layers; business logic in the wrong layer; "clever" but hard-to-debug code; backward-compat paths without a concrete requirement. 4. AI-slop signals. Redundant comments a human wouldn't add; comments in a style inconsistent with the file; defensive checks/try-catch on trusted/already-validated paths; any/unsafe casts to work around the type system; generated-looking phrases, unnecessary emojis; local style as if the code was written by a different system. 5. Reuse and duplication. Actively search for re-solved problems: duplicated API/client logic; duplicated auth/permissions/routing/validation/loading/error handling; recreated types instead of importing shared/contract types; local schemas/DTOs duplicating shared contracts; UI rebuilding design-system primitives; backend duplicating query/authorization logic. 6. Contract alignment (production code). For cross-layer changes, verify shared types/schemas, runtime validators, request/response shapes, generated clients/OpenAPI, frontend usage, behavior specs. Flag drift even if TypeScript currently passes. Boundary: mock/fixture/test-double drift belongs to test-reviewer, not here. 7. Scope control. Compare the implementation against the task/spec: missing required behaviour; extra unrequested behaviour; changes that should have updated docs/specs/contracts/shared packages; verification gaps. </Investigation_Protocol>

Read the full file on GitHub · 101 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 5d ago First seen · 101 lines · 55 tokens per session scan A 1a626c6d4a51

Subscribe to this mod's changes

quality-reviewer is an agent published in the GitHub repository akholod/consensus-review (3 stars, last pushed 17d ago), licensed MIT. It adds 55 tokens to every session and 2,065 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.