CtrlK
BlogDocsLog inGet started
Tessl Logo

moai-domain-humanize

AI text humanization and 윤문 (post-editing) specialist that detects and removes AI tells while preserving meaning, facts, and figures. Covers Korean, English, Japanese, and Chinese with a shared severity model (S1/S2/S3), quality grades (A/B/C/D), and 30%/50% over-editing guardrails. Use to make AI-generated text read as human-authored without changing what it says (de-ai, naturalness pass).

67

Quality

82%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

SecuritybySnyk

Passed

No findings from the security scan

SKILL.md
Quality
Evals
Security

Quality

Content

77%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A well-structured, actionable body with a strong validation-rich workflow and grade/severity machinery, hurt mainly by progressive disclosure: the per-language module files it routes to are referenced but missing from the bundle.

Suggestions

Add the referenced modules/ files (korean.md, english.md, japanese.md, chinese.md, design-copy.md, copy-review.md) to the skill bundle, or remove the routing references if the catalogues are inlined elsewhere.

De-duplicate the guardrail and ledger rules that currently appear in Operating Principles, the dedicated sections, and again inside workflow steps 5 and 6 — consolidate into one authoritative location with cross-references.

Make the change-rate measurement limitation more concise; the current 'Known limitation' parenthetical repeats information already implied by 'LLM estimate'.

DimensionReasoningScore

Conciseness

Dense, information-rich body that avoids explaining concepts Claude already knows, but guardrail and ledger rules are stated in the Operating Principles, restated in dedicated sections, and again in the workflow steps, leaving some tightening on the table.

4 / 5

Actionability

Highly executable guidance with exact severity rules, grade thresholds (0 residual S1, ≤2 residual S2, ≥70%), guardrail cutoffs (>30% WARN, >50% HALT), a 6-point checklist, and an 8-step workflow; the change-rate percentage being an acknowledged LLM estimate is a minor gap.

4 / 5

Workflow Clarity

An explicit 8-step sequenced workflow with validation checkpoints (Invariant Ledger pre-edit, Delta Audit post-edit, rollback on supplied-item violations) and feedback loops (second pass on Grade C, human review on Grade D) for the destructive over-editing surface.

5 / 5

Progressive Disclosure

The body is well-organized with clearly signaled one-level-deep references to six module files (modules/korean.md, english.md, japanese.md, chinese.md, design-copy.md, copy-review.md), but none of those referenced files exist on disk, so the references do not resolve.

3 / 5

Total

16

/

20

Passed

Description

87%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A specific, well-triggered description that concretely states both what the skill does and when to use it, in correct third-person voice. Minor specificity gaps around the prose/copy genre split keep it just short of perfect on the capability dimension.

DimensionReasoningScore

Specificity

Names the domain and several concrete actions ('detects and removes AI tells', 'preserving meaning, facts, and figures', 'make AI-generated text read as human-authored') but does not surface the prose/copy genre modes in the description itself, leaving minor coverage gaps.

4 / 5

Completeness

Clearly answers 'what' (detect/remove AI tells across 4 languages with a shared severity model, grades, and over-editing guardrails) and 'when' with an explicit 'Use to make AI-generated text read as human-authored' trigger clause.

5 / 5

Trigger Term Quality

Includes strong natural terms and synonyms ('AI text humanization', 'post-editing', '윤문', 'AI tells', 'de-ai', 'naturalness pass', 'human-authored'), missing a few common phrasings a user might say like 'make this sound less AI'.

4 / 5

Distinctiveness Conflict Risk

Occupies a clear niche (AI-text humanization / de-ai / 윤문 post-editing) with distinctive multilingual terminology that is unlikely to trigger for the wrong skill.

5 / 5

Total

18

/

20

Passed

Validation

87%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 14 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

allowed_tools_field

'allowed-tools' contains unusual tool name(s)

Warning

frontmatter_unknown_keys

Unknown frontmatter key(s) found; consider removing or moving to metadata

Warning

Total

14

/

16

Passed

Repository
modu-ai/moai-adk
Reviewed

Table of Contents

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.