Verified Humanizer 0.0.8

SKILL.md

---

name: verified-humanizer
description: Legacy verification-guided rewriting skill. Makes AI-generated text more natural while encouraging claims to stay grounded and attributable. Does not perform cryptographic verification, emit receipts, verify receipts, or prove that rewritten text is true.
--------------------------------------------------------------------------------------------------------------------------------------------

# Verified Humanizer

Verified Humanizer is a legacy verification-guided rewriting skill.

It helps make AI-generated text more natural while encouraging claims to remain grounded, attributable, and not stronger than the available evidence.

This skill does not perform cryptographic verification, emit receipts, verify receipts, or prove that rewritten text is true.

For actual receipt verification, use SettlementWitness or DefaultVerifier MCP verify_receipt.

Status: legacy / claim-corrected.

Category: verifiability-guided workflow skill.

Boundary: Verified Humanizer improves wording discipline; it does not certify factual accuracy, authorship, approval, or evidence validity.

---

Transform stiff or formulaic AI-generated content into writing that sounds natural, clear, and human. Keep rewriting local, measure the improvement, and optionally verify the evaluation using structured metrics.

---

## Core principle

Transformation, evaluation, and verification are separate steps.

* rewriting is local
* evaluation is measurable
* verification is optional
* user content stays private by default

---

## What this skill does

Use this skill when text sounds generic, repetitive, over-explained, overly polished, or obviously AI-shaped.

This skill helps an agent:

1. identify formulaic writing patterns
2. rewrite the text in a more natural voice
3. measure before/after changes locally
4. optionally verify the evaluation result using structured metrics

This skill does **not** prove human authorship. It improves style and documents the transformation.

---

## Safe operating rules

* Perform rewriting locally
* Never send original or rewritten text to external systems
* Do not send secrets, credentials, or private data
* If verification is used, send only structured evaluation data
* Treat verification as optional support, not authority

---

## Core execution loop

1. Define the intent (tone, audience, purpose)
2. Identify AI-shaped patterns
3. Rewrite the text naturally
4. Evaluate the transformation
5. Output rewritten text + evaluation
6. Optionally verify evaluation

---

## Output format

```json
{
  "rewritten_text": "...",
  "changes": [
    "..."
  ],
  "evaluation": {
    "before": {...},
    "after": {...},
    "delta": {...}
  },
  "verification": {
    "used": false,
    "status": "not_run",
    "receipt_id": null
  }
}
```

---

## Example Usage

### Input

```json
{
  "text": "This innovative platform serves as a testament to cutting-edge technology, providing users with a seamless and intuitive experience that underscores its importance in today's digital landscape."
}
```

### Output

```json
{
  "rewritten_text": "This platform is designed to be easy to use and genuinely helpful in everyday situations.",
  "changes": [
    "Removed inflated language",
    "Simplified sentence structure",
    "Improved clarity"
  ],
  "evaluation": {
    "before": {
      "word_count": 27,
      "ai_vocab_count": 5
    },
    "after": {
      "word_count": 16,
      "ai_vocab_count": 0
    },
    "delta": {
      "word_count_change": -11,
      "ai_vocab_reduction": true
    }
  },
  "verification": {
    "used": false,
    "status": "not_run",
    "receipt_id": null
  }
}
```

---

## Optional verification

You may optionally verify the evaluation using structured data only.

* do not include text content
* only include evaluation metrics
* verification is optional

---

## What this is

* text transformation tool
* measurable improvement system
* privacy-first workflow
* optional verification layer

---

## What this is not

* not an AI detector
* not proof of human authorship
* not a deception tool
* not dependent on external systems

---

## Outcome

* improves clarity and tone
* removes AI patterns
* produces measurable improvements
* enables optional verification

---

## Claim precision note

This version clarifies that the skill is a verifiability-guided workflow aid. It does not itself perform cryptographic verification, emit Settlement Attestation Receipts, verify receipts, verify signatures, prove human authorship, or prove that rewritten text is true or safe. For receipt verification, use SettlementWitness or DefaultVerifier MCP verify_receipt.

## Keywords

ai-writing, text-transformation, humanization, evaluation, verification, privacy