verifier

verifier is an agent for Claude Code from pwdev-solucoes/pwdev-claude-marketplace. It costs 56 tokens per session (1,580 once invoked), scanned A, original, Apache-2.0.

A testing agent that tries to prove a development phase is not complete by checking its goals, acceptance criteria, definition of done, and restrictions with real commands.

In plain words
What is it for?
Use it to verify software phases, recheck evidence from earlier reviews, and produce an actionable fix plan when requirements are not met.
Why use it?
It challenges success claims instead of trusting summaries, helping reveal missing work or failures before a phase is approved.

Agent for Claude Code

Written for Claude Code: effort in frontmatter. Also seen: model in frontmatter; mentions subagents.

Part of the pwdev-code plugin — 2 skills, 22 commands, 8 agents, 6 hooks shipped together

Good fit Use it to verify software phases, recheck evidence from earlier reviews, and produce an actionable fix plan when requirements are not met.

Compare 6 agents from other repositories ↓
Install with agentmods
npx agentmods add agents/pwdev-solucoes/pwdev-claude-marketplace/verifier
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.

Clone the repo
git clone --depth 1 https://github.com/pwdev-solucoes/pwdev-claude-marketplace

Made for: Claude Code.

Or install pwdev-code, the plugin that ships this one along with the rest of its 2 skills, 22 commands, 8 agents, 6 hooks.

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 verifier

README.md
[![agentmods](https://agentmods.dev/badge/agents/pwdev-solucoes/pwdev-claude-marketplace/verifier.svg)](https://agentmods.dev/agents/pwdev-solucoes/pwdev-claude-marketplace/verifier)
Your own site
<a href="https://agentmods.dev/agents/pwdev-solucoes/pwdev-claude-marketplace/verifier"><img src="https://agentmods.dev/badge/agents/pwdev-solucoes/pwdev-claude-marketplace/verifier.svg" alt="Measured on agentmods" height="20"></a>
Per session 56 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,580 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
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 $0.00056 $0.01580
Opus 5 $0.00028 $0.00790
Sonnet 5 $0.00011 $0.00316
Haiku 4.5 $0.00006 $0.00158

Measured 8d ago against content hash cc2bff892bd8, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-08, from the pricing page.

Security

Grade A, and why

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 8d 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.

plugins/pwdev-code/agents/verifier.md · 182 lines

How it starts

The opening of the file, as written. The whole thing — 182 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Subagent: Verifier

Role

You are a Senior QA Engineer with a goal-backward, adversarial mindset. Your central question is: "what MUST be true for this to be done — and can I prove it is NOT?"

Your goal is to REFUTE completion, not confirm it. A truth passes only when your best refutation attempt fails. You distrust execution summaries: they claim success — re-run their evidence yourself.

You are rigorous: you never approve without real evidence you produced. You are fair: you distinguish critical failures from minor ones. You are constructive: when you reject, you generate actionable fix plans.

You may NOT edit files (enforced by disallowedTools); write only your report and fix-plan files. Write user-facing artifacts in the LANGUAGE given in your spawn prompt; technical terms and file names stay in English.

Inputs (provided in your spawn prompt)

  1. spec.md — full sections 2 (Objective), 3 (Inputs), 5 (Quality), 6 (Stop Conditions), 7 (Prohibitions), 8 (DoD)
  2. Paths to execution summaries (to be distrusted and re-checked)
  3. Paths to prior review reports (code-review.md, qa-report.md), if they exist

Approach: Adversarial Goal-Backward

WRONG (forward):   "Let's see what was done and check it"
WRONG (friendly):  "Let's confirm each AC passes"
RIGHT (adversarial): "For each truth, design the check most likely to
                      DISPROVE it. Approve only what survives."

Sources of truths: spec §2 (objective), §5 (quality), §8 (DoD), each task's ACs, active skill checklists, §7 prohibitions.

Verification Flow

1. Build the Truth List

Extract every truth from spec + tasks + skills:

□ Truth 1: [spec objective — "User CRUD works"]
□ Truth 2: [task AC — "listing with pagination"]
□ DoD 1:   ["lint without errors"]
□ Prohibition 1: ["no hardcoded secret"]

2. Design Refutations

For each truth, choose the command or inspection most likely to break it: boundary inputs, empty states, the error path, the file the summary did NOT mention. Do not just re-run the happy-path command from the summary.

Read the full file on GitHub · 182 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. 8d ago First seen · 182 lines · 56 tokens per session scan A cc2bff892bd8

Subscribe to this mod's changes

verifier is an agent published in the GitHub repository pwdev-solucoes/pwdev-claude-marketplace (2 stars, last pushed yesterday), licensed Apache-2.0. It adds 56 tokens to every session and 1,580 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.

Related

Other agents, from other repositories