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.
86
78%
Does it follow best practices?
Impact
100%
1.09xAverage score across 3 eval scenarios
Passed
No known issues
Optimize this skill with Tessl
npx tessl skill review --optimize ./prompts/humanizer/SKILL.mdQuality
Discovery
100%Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.
This is an excellent skill description that clearly communicates its purpose, provides explicit trigger guidance, and lists specific capabilities in detail. It uses proper third-person voice, includes a 'Use when' clause, and is distinctive enough to avoid conflicts with general writing or editing skills. The enumeration of specific AI writing patterns adds both specificity and searchability.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Lists multiple specific concrete actions and patterns: inflated symbolism, promotional language, superficial -ing analyses, vague attributions, em dash overuse, rule of three, AI vocabulary words, negative parallelisms, and excessive conjunctive phrases. Very detailed about what it detects and fixes. | 3 / 3 |
Completeness | Clearly answers both what ('Remove signs of AI-generated writing', 'Detects and fixes patterns including...') and when ('Use when editing or reviewing text to make it sound more natural and human-written') with an explicit 'Use when' clause. | 3 / 3 |
Trigger Term Quality | Includes strong natural trigger terms users would say: 'AI-generated writing', 'sound more natural', 'human-written', 'editing', 'reviewing text'. Also includes specific pattern names that knowledgeable users might reference like 'em dash overuse' and 'AI vocabulary words'. | 3 / 3 |
Distinctiveness Conflict Risk | Occupies a very clear niche — specifically about detecting and removing AI writing patterns. The reference to Wikipedia's guide and the enumerated list of specific AI writing patterns make this highly distinctive and unlikely to conflict with general editing or writing skills. | 3 / 3 |
Total | 12 / 12 Passed |
Implementation
57%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This skill is exceptionally thorough and actionable, providing concrete before/after examples and word-watch lists for 24 distinct AI writing patterns. Its main weakness is length—it's a monolithic document that could benefit significantly from splitting pattern catalogs into separate reference files. The workflow is functional but could use stronger validation steps and prioritization guidance.
Suggestions
Split the 24 patterns into separate reference files (e.g., content_patterns.md, language_patterns.md, style_patterns.md, communication_patterns.md) and keep SKILL.md as a concise overview with the process, key principles, and links to pattern catalogs.
Add prioritization guidance to the workflow—when text has 10+ patterns, which should be fixed first? Some patterns (vague attributions, promotional language) change meaning more than others (em dashes, title case).
Condense simpler patterns (curly quotes, emojis, title case, boldface overuse) into a single 'quick fixes' table rather than giving each a full section with before/after examples.
Add a validation step to the workflow: after rewriting, check that no new AI patterns were introduced and that factual claims from the original are preserved or flagged as unverifiable.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The skill is comprehensive and mostly well-organized, but at ~400+ lines it's quite long. Many of the 24 patterns could be condensed—Claude doesn't need extensive before/after examples for every single pattern. The 'Personality and Soul' section, while valuable, adds significant length. Some patterns (curly quotes, emojis, title case) could be single-line entries. | 2 / 3 |
Actionability | Highly actionable with concrete before/after examples for every pattern, specific words-to-watch lists, and a clear process. Claude can directly apply each pattern identification and rewrite technique. The full example at the end ties everything together with annotated changes. | 3 / 3 |
Workflow Clarity | The 5-step process at the end is clear but basic—it lacks validation checkpoints. There's no guidance on how to handle conflicts between patterns, how to prioritize when text has many issues, or how to verify the output doesn't introduce new problems. For a text transformation task, a 'read it aloud' check is mentioned but not structured as a validation gate. | 2 / 3 |
Progressive Disclosure | This is a monolithic wall of content—all 24 patterns with full examples are inline in a single file. The patterns could easily be split into separate reference files (content_patterns.md, language_patterns.md, style_patterns.md) with SKILL.md serving as an overview with the process and key principles. The reference link at the bottom is good but doesn't compensate for the lack of internal structure. | 1 / 3 |
Total | 8 / 12 Passed |
Validation
100%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 11 / 11 Passed
Validation for skill structure
No warnings or errors.
- Repository
- jdrhyne/agent-skills
- Commit
6768672
- 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.