humanizer

humanizer is an agent for Claude Code from XuanRanL/loamwright-SEO-Skill. It costs 60 tokens per session (2,341 once invoked), scanned A, original, Apache-2.0.

A draft editor that detects common AI writing patterns and revises text to match a chosen voice and purpose. It supports detection, rewriting, and editing modes, with reports about the changes.

In plain words
What is it for?
Use it to inspect or revise an existing draft, apply a professional or other configured voice, run writing-quality checks, and create a humanizer report.
Why use it?
It helps make generated drafts sound less formulaic while preserving human edits and recording the checks it performed.

Agent for Claude Code

Written for Claude Code: shipped in a Claude Code plugin. Also seen: model in frontmatter; mentions subagents.

Part of the xuanran-seo-blog-writer plugin — 68 skills, 34 agents, 4 hooks 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/xuanranl/loamwright-seo-skill/humanizer
Clone the repo
git clone --depth 1 https://github.com/XuanRanL/loamwright-SEO-Skill

Made for: Claude Code.

Or install xuanran-seo-blog-writer, the plugin that ships this one along with the rest of its 68 skills, 34 agents, 4 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 humanizer

README.md
[![agentmods](https://agentmods.dev/badge/agents/xuanranl/loamwright-seo-skill/humanizer.svg)](https://agentmods.dev/agents/xuanranl/loamwright-seo-skill/humanizer)
Your own site
<a href="https://agentmods.dev/agents/xuanranl/loamwright-seo-skill/humanizer"><img src="https://agentmods.dev/badge/agents/xuanranl/loamwright-seo-skill/humanizer.svg" alt="Measured on agentmods" height="20"></a>
Per session 60 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 2,341 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 $0.00060 $0.02341
Opus 5 $0.00030 $0.01171
Sonnet 5 $0.00012 $0.00468
Haiku 4.5 $0.00006 $0.00234

Measured 6d ago against content hash 09a584acc81f, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-06, from the pricing page.

Security

Grade A, and why

humanizer 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 6d 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/humanizer.md · 180 lines

How it starts

The opening of the file, as written. The whole thing — 180 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Humanizer Agent

You remove AI fingerprints from drafted content. You apply the brand's voice + purpose pair (default: professional × general).

Physical constraint

Tool whitelist is Read + Edit + Write + Bash. Use Edit for draft.md edits (preserves human edits). Use Write ONLY for creating humanizer-report.json (your output artifact). Use Bash to run lint scripts (banned_word_lint.py, ai_tells_detector.py) for detection.

CRITICAL: Always write humanizer-report.json before finishing

Even if you run out of iterations or encounter issues, you MUST write humanizer-report.json to the workspace as your LAST action. An incomplete report with "_generated_by": "humanizer-subagent" is far better than no report — the pipeline blocks without this artifact.

MANDATORY output — humanizer-report.json

You MUST Write memory/workspace/{task}/humanizer-report.json before your last turn. If you run out of iterations, write it with your best current results. An incomplete report is better than no report — the pre-publish gate hard-blocks on missing humanizer-report.json.

{
  "ai_slop_score": 14,
  "iterations": 1,
  "patterns_found": {},
  "patterns_fixed": {},
  "scaffold_markers_removed": [],
  "_generated_by": "humanizer-subagent"
}

Inputs

  • target_file (typically memory/workspace/{task}/draft.md)
  • voice (e.g. "professional", "casual", "warm", "blunt", "technical")
  • purpose (e.g. "general", "marketing", "essay", "technical", "email")
  • mode: "detect" | "rewrite" | "edit"
    • detect: read-only audit, output report
    • rewrite: regenerate full sections (used when surgical edits insufficient)
    • edit: surgical line-by-line fixes (default; preserves the rest)
  • score_target: pass when AI-Slop score < this (default 20)
  • max_iterations: cap iterations (default 3)

References you load

CRITICAL — load first, every run:

  • references/style/markdown-authoring-conventions.md — 6 publish-blocking rules (raw HTML escapes, Pandoc anchors, hand-rolled srcset, image placeholder syntax). Any rewrite that introduces a violation here will be hard-veto'd at render_lint. Re-check after every edit pass that you have not introduced raw <strong> / <em> / <p> / <ol> / <a target> / <img srcset> / {#anchor}.

Read the full file on GitHub · 180 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. 6d ago First seen · 180 lines · 60 tokens per session scan A 09a584acc81f

Subscribe to this mod's changes

humanizer is an agent published in the GitHub repository XuanRanL/loamwright-SEO-Skill (47 stars, last pushed 19d ago), licensed Apache-2.0. It adds 60 tokens to every session and 2,341 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 2026-08-30.

Related

Other agents, from other repositories

ai-search-geo-specialist

Read-only AI-search (GEO/AEO) specialist. Use proactively during an audit to evaluate answer extractability/passage structure, fact density and original data, AI-crawler access and citability, llms.txt, and entity/knowledge-graph linkage.

Hainrixz/claude-seo-ai · 59 tokens

schema-generator

Structured-data specialist. Use proactively during an audit to validate existing JSON-LD and PROPOSE complete Tier-1 schema blocks. It proposes diffs only and does NOT write files.

Hainrixz/claude-seo-ai · 39 tokens

technical-auditor

Read-only technical SEO specialist. Use proactively during an audit to analyze crawlability, indexability, rendering, Core Web Vitals, mobile-friendliness, title/meta/head hygiene, heading structure, social cards, images, internal linking, and sitemaps.

Hainrixz/claude-seo-ai · 57 tokens

content-eeat-analyst

Read-only content quality specialist. Use proactively during an audit to evaluate E-E-A-T (author identity, credentials, trust signals, transparency) and content freshness/temporal signals.

Hainrixz/claude-seo-ai · 43 tokens

seo-fixer-writer

The ONLY agent allowed to write files. Used exclusively by the fix skill (the /claude-seo-ai:fix command) AFTER the user has confirmed the diffs. Applies AUTO-class fixes, backs up first, is idempotent, git-aware, and re-verifies each change.

Hainrixz/claude-seo-ai · 65 tokens

geo-schema

Schema markup specialist detecting, validating, and generating structured data (JSON-LD preferred). Focuses on schemas that improve AI discoverability including Organization, Person, Article, sameAs, and speakable properties.

Cognitic-Labs/geoskills · 44 tokens