Humanize Text

PassAudited by VirusTotal on May 11, 2026.

Overview

Type: OpenClaw Skill Name: humanize-text Version: 1.0.0 The skill bundle is classified as suspicious due to its explicit instructions for file read/write operations in SKILL.md and its implementation in scripts/humanize.py. While these capabilities are plausibly needed for the stated purpose of humanizing text, they represent a significant 'risky capability' as defined, allowing the agent to modify arbitrary files it is instructed to process. There is no evidence of intentional malicious behavior such as data exfiltration or persistence, and the SKILL.md includes guardrails against altering meaning or technical accuracy. However, the extensive and complex prompt-injection instructions in SKILL.md, combined with direct file system access, elevate the risk beyond benign, as an agent could potentially be prompted to apply these transformations to sensitive system files.

Findings (0)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

What this means

A selected document could be changed in ways that alter wording or nuance.

Why it was flagged

The skill may cause the agent to modify user files. This is aligned with a text-editing skill, but users should review changes before overwriting important content.

Skill content
For files, read the file, apply transformations, write it back.
Recommendation

Use the diff-summary step, keep backups for important files, and prefer approval before overwriting originals.

What this means

Readers or reviewers may assume the text was written without AI assistance.

Why it was flagged

The skill is explicitly intended to make AI-assisted text appear more naturally human-written. This is disclosed and purpose-aligned, but may matter in contexts with AI-use disclosure rules.

Skill content
Humanize AI-generated text by removing telltale AI writing patterns... make it read like a human wrote it.
Recommendation

Use only where this kind of editing is allowed, and follow any disclosure, academic, workplace, or platform policies that apply.