humanizer

Remove signs of AI-generated writing from text. Use when editing or reviewing text to make it sound more natural and human-written. Based on Wikipedia's comprehensive "Signs of AI writing" guide. Detects and fixes patterns including inflated symbolism, promotional language, superficial -ing analyses, vague attributions, em dash overuse, rule of three, AI vocabulary words, negative parallelisms, and excessive conjunctive phrases.

1.09x

Quality

73%

Does it follow best practices?

Impact

100%

1.09x

Average score across 3 eval scenarios

Securityby

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./prompts/humanizer/SKILL.md

The canonical home for this skill is humanizer in jdrhyne/agent-skills

Quality

Content

63%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 thorough, well-structured catalog of AI-writing patterns with concrete before/after examples that make it highly actionable. Its main weaknesses are length/verbosity across 24 near-parallel sections and a single-file monolithic structure that forgoes progressive disclosure via reference files.

Suggestions

Move the bulk of the 24-pattern catalog into a references file (e.g. PATTERNS.md) and keep SKILL.md as an overview with a short quick-start and a pointer, improving token efficiency and navigation.

Tighten repetitive pattern sections: each already follows the same structure, so the prose explanations can be shortened once the format is established.

Add an explicit verification checkpoint in the Process section (e.g. 're-read aloud and confirm no listed trigger words remain') to strengthen the feedback loop.

Dimension	Reasoning	Score
Conciseness	The document is comprehensive but lengthy (~420 lines) with many before/after blocks that, while illustrative, repeat the same core instruction across 24 patterns; some explanatory prose ('LLMs use statistical algorithms...', the full soul/voice section) could be trimmed. It is mostly efficient but could be tightened.	3 / 5
Actionability	Each pattern gives concrete 'Words to watch', a stated problem, and concrete before/after rewrites that model executable fixes, with minor gaps where no explicit word list is given (e.g. negative parallelisms, rule of three).	4 / 5
Workflow Clarity	A clear sequenced 'Process' section (read, identify, rewrite, verify naturalness, present) with verification criteria, though the validation is qualitative ('sounds natural when read aloud') rather than a hard checkpoint, and the skill is non-destructive so the destructive/batch cap does not apply.	4 / 5
Progressive Disclosure	Content is well-organized by section headings but everything is inlined into a single monolithic SKILL.md with no bundle files or one-level-deep references; the large pattern catalog could be split into a separate reference file for easier navigation.	3 / 5
	Total	14 / 20 Passed

Description

83%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 strong, third-person description that clearly states both the skill's function and its use conditions, anchored to a named source (Wikipedia's 'Signs of AI writing'). It enumerates concrete patterns, making it largely comprehensive and distinct, with only minor keyword-coverage gaps.

Suggestions

Add common synonyms users might say, e.g. 'de-AI text', 'AI-sounding writing', or 'humanize text', to broaden natural trigger coverage.

Tighten the trigger clause by specifying a domain ('Use when editing prose...') to reduce minor overlap with general editing skills.

Dimension	Reasoning	Score
Specificity	Lists several concrete actions ('Detects and fixes patterns', 'Remove signs of AI-generated writing') and enumerates the specific pattern categories (inflated symbolism, promotional language, rule of three, etc.), with only minor gaps in coverage of how fixes are applied.	4 / 5
Completeness	Explicitly answers both 'what' (remove signs of AI-generated writing, detect and fix named patterns) and 'when' ('Use when editing or reviewing text to make it sound more natural and human-written') with concrete trigger phrasing.	5 / 5
Trigger Term Quality	Includes natural phrases users would say ('editing or reviewing text', 'make it sound more natural and human-written') alongside the domain term 'AI-generated writing', though it omits common synonyms like 'AI-sounding' or 'de-AI'.	4 / 5
Distinctiveness Conflict Risk	Targets a clear niche (AI-writing-pattern detection/removal) with distinct triggers, though the broad framing of 'editing or reviewing text' could create minor overlap with general editing skills.	4 / 5
	Total	17 / 20 Passed

Validation

100%

Warnings & errors only

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

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: bap-jorkim/agent-skills-fork-feb-25
Path: prompts/humanizer/SKILL.md
Commit: f4ed46f

Reviewed: 2 days ago

Table of Contents

Discovery Implementation Validation

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.