Humanizer
ReviewAudited by ClawScan on May 10, 2026.
Overview
Prompt-injection indicators were detected in the submitted artifacts (system-prompt-override); human review is required before treating this skill as clean.
This appears safe to use as a writing-editing skill if you want its humanization behavior. Be careful with the optional always-on system prompt because it changes general agent style, and run any API/MCP integrations locally or in a trusted environment when working with private documents. ClawScan detected prompt-injection indicators (system-prompt-override), so this skill requires review even though the model response was benign.
Findings (3)
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.
If copied into a system prompt, the writing rules may affect all future responses, not just explicitly humanized drafts.
This is user-directed guidance for changing an agent's default writing behavior across future interactions. It is purpose-aligned, but users should only apply it if they want the skill's style rules always active.
For LLMs that should ALWAYS write like a human ... Add to your project instructions or custom system prompt:
Use the always-on instructions only in projects where that global writing style is desired; otherwise invoke the skill only for specific text-editing tasks.
Private text sent to a deployed API or connected action may be exposed to that server environment.
The optional HTTP API accepts text for scoring, analysis, humanization, and stats, and allows browser access from any origin. No storage or exfiltration is shown, but deployment choices affect privacy for submitted drafts.
'Access-Control-Allow-Origin': '*'
Run the API locally or behind access controls for confidential drafts, and only connect OpenAI Actions or other clients to deployments you trust.
Manual MCP setup depends on the npm package supply chain rather than only the instruction-only skill file.
The optional MCP server depends on an external package using a semver range. This is normal for an MCP integration, but it means manual npm installs may pull later compatible versions.
"@modelcontextprotocol/sdk": "^1.0.0"
Review dependencies and consider pinning versions or using a lockfile before deploying the optional MCP server in sensitive environments.