agentic-humanizer

Agentic Humanizer

A 5-pass AI humanizer. It always runs the core rewrite workflow; Slop or Not Pro adds measured on-device AI detector checks.

• Without Slop or Not: runs the full rewrite workflow.
• With Slop or Not Pro: adds on-device AI detector scoring, readability, Text Cleanup, and cleanup stats.

Slash command: /agentic-humanizer [paste text]

Inline overrides: /agentic-humanizer dialect=us|uk grade=N tone=casual|professional|academic length=±10|exp|trim threshold=N max=N voice=/path/to/file.txt|off voice-skip skip-interview [paste]

What this skill does

1. Detects the host harness (Claude Code, Codex, Cursor, Gemini CLI, OpenCode, or generic).
2. Handles profile and voice management commands before any rewrite.
3. Resolves rewrite preferences from inline overrides, saved profile, or the harness interview.
4. Optionally resolves a writing sample and extracts a cached stylometric fingerprint. Voice matching does not require Slop or Not.
5. Probes whether Slop or Not Pro is reachable via MCP or CLI.
6. Runs the 5-pass humanization workflow:
   • Core mode logs unscored iterations.
   • Slop or Not Pro runs Text Cleanup, detection, and readability checks.
7. Returns the final text, loop history, highest-impact edits, and, when Slop cleanup ran, a Text Cleanup summary.

Step 1: Detect the harness

Identify which harness is running by checking for the harness's distinctive question tool. Use the first match:

| Harness | Distinctive tool present? | Read this file |

Step 2: Profile management commands

Step 3: Resolve rewrite preferences

Step 4: Resolve voice sample

Step 5: Probe Slop or Not Pro

Step 6: Run the loop

Slop or Not Pro setup

Core setup

Cleanup stats parsing

Termination with Slop or Not Pro

Completion in Core mode

Mid-flight Pro-gate