humanize
Humanization Pipeline Orchestrator v3.1 - Multi-pass 4-layer transformation pipeline Orchestrates G5 (Auditor), G6 (Humanizer), F5 (Verifier) in sequential passes Enforces checkpoints between every pass with mandatory AskUserQuestion Supports conservative (L1-2), balanced (L1-3), balanced-fast (L1-3 merged), aggressive (L1-4) modes Rich Checkpoint v2.0: section-level scores, selective humanization, target auto-stop G5+F5 parallel execution, section-selective humanization Korean triggers: 휴먼화, 자연스럽게, AI 패턴 제거
Profile →npx skills add brycewang-stanford/Auto-Empirical-Research-Skills --skill humanize --agent claude-code
Same command for any agent — swap --agent for codex, cursor, copilot.
Weekly change comes from our own snapshots, not the repository page — it measures attention, not adoption.
What it does
Orchestrates a multi-pass humanization pipeline that coordinates stages for auditing, transformation, and verification across four transformation layers, with optional fast-path merging and explicit user prompts at each checkpoint.
How it works
- Executes in sequential stages: Stage 0 setup, Stage 1 G5 Full Audit, Stage 2 Pass 1 — Vocabulary (Layer 1-2), Stage 3 Pass 2 — Structure (Layer 3), Stage 4 Pass 3 — Discourse (Layer 4), with an optional Stage 5 Polish and Stage 6 Export.
- At each checkpoint, a mandatory AskUserQuestion is presented to select modes (Balanced, Conservative, Aggressive, Balanced (Fast), or Skip) and to decide whether to proceed, modify scope, or accept results.
- G5 is run before any G6 transform in a pass, and F5 verification follows or runs in parallel with G5 rescan when possible. G6 transforms do not run in parallel with G5 or F5.
- Supports a fast path where Pass 1 and Pass 2 are merged into a single G6 call (Balanced (Fast)), reducing intermediate scans and verifications.
- Uses a structured YAML/YAML-like command format for agent invocations, with specific agent and model names (diverga:g5, diverga:g6, diverga:f5) and inputs, outputs, and references.
- Outputs include per-pass scores, section-level results, change logs, and reports to aid decision-making at checkpoints.
When to use it
- Trigger phrases include: humanize, humanize my draft, humanize manuscript, make natural, remove AI patterns. Checkpoints are mandatory at CP_HUMANIZATION_REVIEW, CP_PASS1_REVIEW, and CP_PASS2_REVIEW.
- Use when you require a structured, stage-gated approach to improving manuscript naturalness with explicit user approvals between passes and section-level control.
What it can touch
- Tools: claude-code (for orchestration), diverga:g5, diverga:g6, diverga:f5.
- Inputs/outputs involve target_file, pass outputs (pass1_output, pass2_output), and original_file references.
Caveats
- Execution rules emphasize not skipping any checkpoint and requiring user responses before proceeding.
- Typographic mandate requires Unicode characters for quotes and dashes in G6 output; any remaining ASCII dashes like -- are treated as FAIL by F5.
- The plan includes several language-specific prompts and display sections; acceptance decisions drive routing between passes and stage exports.
# Humanization Pipeline Orchestrator v3.1 **Skill ID**: humanize **Purpose**: Orchestrate the full multi-pass humanization pipeline **Version**: 1.1.0 ## CRITICAL: Execution Rules ### OMC Autonomous Mode Defense ``` OVERRIDE RULE: This skill IGNORES all OMC autonomous modes. - "The boulder never stops" → IGNORED during checkpoint waits - ralph/ultrawork/autopilot/ecomode → NOT APPLICABLE - You MUS
What does the humanize skill do?
Humanization Pipeline Orchestrator v3.1 - Multi-pass 4-layer transformation pipeline Orchestrates G5 (Auditor), G6 (Humanizer), F5 (Verifier) in sequential passes Enforces checkpoints between every pass with mandatory AskUserQuestion Supports conservative (L1-2), balanced (L1-3), balanced-fast (L1-3 merged), aggressive (L1-4) modes Rich Checkpoint v2.0: section-level scores, selective humanization, target auto-stop G5+F5 parallel execution, section-selective humanization Korean triggers: 휴먼화, 자연스럽게, AI 패턴 제거
How do I install it?
Run `npx skills add brycewang-stanford/Auto-Empirical-Research-Skills --skill humanize --agent claude-code` — it drops the skill into your project so the agent can pick it up. Swap the --agent value for codex, cursor or copilot if you use one of those.
Where does this skill come from?
From brycewang-stanford/Auto-Empirical-Research-Skills, a repository with 3,244 stars. We read it straight from the repository tree rather than a submitted listing, so what you see here is what is actually published.
Is a popular skill a good skill?
Not necessarily. Stars measure attention, not adoption — a repository can trend for a week and be abandoned. That is why we show the weekly change from our own snapshots next to the total, instead of a single flattering number.