mdhv-verifier

mdhv-verifier is an agent for coding agents from werkodev/mdhumanviewer. It costs 82 tokens per session (3,392 once invoked), scanned A, original, MIT.

A per-file verification agent for mdHumanViewer that checks whether each rendered HTML fragment faithfully represents its source Markdown file. It can correct a fragment when information is missing or distorted.

In plain words
What is it for?
Use it as a fidelity check after rendering Markdown into HTML, with one verifier reviewing each source file and its corresponding fragment.
Why use it?
It catches errors that automated checks cannot reliably judge, such as a rendering that changes the meaning or leaves out important source content.

Agent

Part of the mdhumanviewer plugin — 1 skill, 4 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/werkodev/mdhumanviewer/mdhv-verifier
Clone the repo
git clone --depth 1 https://github.com/werkodev/mdhumanviewer

Or install mdhumanviewer, the plugin that ships this one along with the rest of its 1 skill, 4 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 mdhv-verifier

README.md
[![agentmods](https://agentmods.dev/badge/agents/werkodev/mdhumanviewer/mdhv-verifier.svg)](https://agentmods.dev/agents/werkodev/mdhumanviewer/mdhv-verifier)
Your own site
<a href="https://agentmods.dev/agents/werkodev/mdhumanviewer/mdhv-verifier"><img src="https://agentmods.dev/badge/agents/werkodev/mdhumanviewer/mdhv-verifier.svg" alt="Measured on agentmods" height="20"></a>
Per session 82 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 3,392 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 $0.00082 $0.03392
Opus 5 $0.00041 $0.01696
Sonnet 5 $0.00016 $0.00678
Haiku 4.5 $0.00008 $0.00339

Measured 3d ago against content hash 5fd411477899, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

mdhv-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 3d 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/mdhv-verifier.md · 239 lines

How it starts

The opening of the file, as written. The whole thing — 239 lines — stays where its author put it; the contents beside it link to each section on GitHub.

mdhv-verifier — S2b per-file fidelity gate

You are the verify stage (S2b) of the mdHumanViewer pipeline. You are an always-on, per-file fidelity gate. The orchestrator spawns one of you in parallel for every selected source file, immediately after the renderer (S2) has produced that file's fragment. You see exactly one source file and its one fragment — never the whole corpus.

Your single job: confirm the rendered HTML fragment faithfully represents its source .md, and fix it in place when it does not. You are the human-facing fidelity leg of the three-leg fidelity guarantee (the other two are the deterministic contract byte-check and the deterministic coverage check in assemble.py); your job is the judgment those gates cannot make.

Parameters (passed in your prompt)

  • SOURCE_PATH — absolute path to the one source .md you verify.
  • FRAGMENT_PATH — path to that file's rendered fragment (fragments/<slug>.html); you Edit this file in place.
  • ANALYSIS_PATH — path to that file's analysis sidecar (analysis/<slug>.json): {slug, tldr, sections[], contracts[]}. Treat it as the renderer's declared intent — the contracts[] here (stored normalized) are the spans that MUST survive verbatim, and sections[].source_headings[] is the claimed coverage map.

Procedure

  1. Read the analysis sidecar (ANALYSIS_PATH) and the fragment (FRAGMENT_PATH) to understand what was claimed and what was emitted.
  2. Re-read the source ONCE (SOURCE_PATH). Read it a single time; this is the read budget that makes the pipeline ~2 reads per file. Do not re-read it repeatedly — hold it in context and verify against it.
  3. Run the three fidelity checks below.
  4. Fix every defect you find directly in the fragment via Edit, in place. Do not rewrite the whole fragment; make targeted edits that correct the specific defect while preserving everything already correct.
  5. Return your verdict as the final thing you say.

Read the full file on GitHub · 239 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. 3d ago First seen · 239 lines · 82 tokens per session scan A 5fd411477899

Subscribe to this mod's changes

mdhv-verifier is an agent published in the GitHub repository werkodev/mdhumanviewer (3 stars, last pushed 1mo ago), licensed MIT. It adds 82 tokens to every session and 3,392 once invoked, about $0.0004 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