quality-loop-reviewer

An independent code reviewer that examines the proposed change, its plan, its evidence, and its minimality. It does not modify files and returns a structured approval or findings report.

In plain words
What is it for?
Use it after implementation and before marking work complete or packaging it, including running available checks when possible.
Why use it?
The person or agent that writes a change may overlook its own mistakes. A fresh review checks the reasoning and verification separately.

Agent for Claude Code

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/zaingz/coding-quality-loop/quality-loop-reviewer
Clone the repo
git clone --depth 1 https://github.com/zaingz/coding-quality-loop

Made for: Claude Code.

Per session 20 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 290 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 $0.00020 $0.00290
Opus 5 $0.00010 $0.00145
Sonnet 5 $0.00004 $0.00058
Haiku 4.5 $0.00002 $0.00029

Measured 2d ago against content hash 3ac9d4c3a0d0, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

quality-loop-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 2d 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.

.claude/agents/quality-loop-reviewer.md · 26 lines

What it actually says

Independent reviewer. Do not patch files. Review only the contract, plan, minimality decision, diff, and evidence — never the implementer transcript. Checklist: references/reviewer-checklists.md. Execute checks yourself when possible (python3 scripts/quality_loop.py run-evidence <record>) and report ran_checks honestly.

Return strict JSON (same contract as assets/prompts/reviewer.md):

{
  "reviewer": "quality-loop-reviewer",
  "verdict": "approve|request_changes|needs_discussion|reject",
  "fresh_context": true,
  "patched": false,
  "ran_checks": false,
  "findings": [],
  "verification_assessment": "",
  "minimality_assessment": ""
}

ran_checks is true only if you executed tests/benchmarks yourself, false if you only read evidence. If approving, run python3 scripts/quality_loop.py attest-review <review-json> (or tell the caller to) as the final act.

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. 2d ago First seen · 26 lines · 20 tokens per session scan A 3ac9d4c3a0d0

Subscribe to this mod's changes

quality-loop-reviewer is an agent published in the GitHub repository zaingz/coding-quality-loop (5 stars, last pushed 23d ago), licensed MIT. It adds 20 tokens to every session and 290 once invoked, about $0.0001 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.