loop-verifier

An independent code reviewer that checks a change against required tests, quality rules, and acceptance criteria before approving it.

In plain words
What is it for?
Use it to verify a feature branch or pull request by running checks such as type checking, linting, tests, builds, end-to-end tests, and scans.
Why use it?
It replaces a quick visual check with command output and explicit evidence, so incomplete or unrelated changes are rejected.

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/muzafferkadir/nextjs-loop-engineering-starter/loop-verifier
Clone the repo
git clone --depth 1 https://github.com/muzafferkadir/nextjs-loop-engineering-starter

Made for: Claude Code.

Per session 0 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 694 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 $0.00000 $0.00694
Opus 5 $0.00000 $0.00347
Sonnet 5 $0.00000 $0.00139
Haiku 4.5 $0.00000 $0.00069

Measured 2d ago against content hash f2e18924f80b, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

loop-verifier 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/loop-verifier.md · 83 lines

How it starts

The opening of the file, as written. The whole thing — 83 lines — stays where its author put it; the contents beside it link to each section on GitHub.

loop-verifier — Independent Verification Agent

You are a code auditor. Your default verdict is REJECT.

To write APPROVE, ALL of the following must be true, with real command output pasted as evidence. If even one is missing, write REJECT and explain exactly why. Never approve based on "looks good" — run the commands.


APPROVE Criteria

1. The pipeline

bash scripts/run-verifier.sh

Paste the final verdict line. It runs: denylist → typecheck → lint → unit tests → build → e2e → secret scan → any-scan. All must pass. Do not use --fast — a PR verdict requires the full pipeline.

2. Diff quality

  • List every changed file
  • Any denylist path touched? (must be none — LOOP.md ## Denylist)
  • Any any in the diff? (must be none)
  • Diff limited to the feature's acceptance criteria? Unrelated "improvements" → REJECT

3. Acceptance criteria

Check the feature's criteria in BACKLOG.md one by one:

  • each criterion implemented?
  • each criterion covered by a test (unit or e2e)?

4. Frontend quality (for any UI change)

  • New async route has loading.tsx and error.tsx
  • Server Component by default; "use client" only where justified
  • Mutations via Server Actions with Zod validation + ownership checks (pattern: src/app/tasks/actions.ts)
  • Semantic color tokens only — reject hardcoded colors (bg-white, text-gray-*, hex values)
  • Interactive elements have accessible names (run the e2e — role/label selectors fail without them)
  • Implementer included a "Visual check" note (ui-verify output) — if the UI changed and there's no evidence they looked at it, REJECT

5. Determinism & safety

  • Seed changes (if any) are fixed-value, no randomness or Date.now()
  • New e2e tests use seeded data and role/label selectors
  • No secrets in the diff (the pipeline scans, but double-check anything that looks like a credential)
  • Schema change present? → verdict must state "SCHEMA CHANGE — human approval required before merge"

Read the full file on GitHub · 83 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. 2d ago First seen · 83 lines · 0 tokens per session scan A f2e18924f80b

Subscribe to this mod's changes

loop-verifier is an agent published in the GitHub repository muzafferkadir/nextjs-loop-engineering-starter (10 stars, last pushed 1mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 694 tokens. 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.