| 100.00% | Adverbs in dialogue tags | Target: ≤10% dialogue tags with adverbs | | totalTags | 11 | | adverbTagCount | 0 | | adverbTags | (empty) | | dialogueSentences | 27 | | tagDensity | 0.407 | | leniency | 0.815 | | rawRatio | 0 | | effectiveRatio | 0 | |
| 92.97% | AI-ism adverb frequency | Target: <2% AI-ism adverbs (58 tracked) | | wordCount | 711 | | totalAiIsmAdverbs | 1 | | found | | | highlights | | |
| 100.00% | AI-ism character names | Target: 0 AI-default names (17 tracked, −20% each) | | codexExemptions | (empty) | | found | (empty) | |
| 100.00% | AI-ism location names | Target: 0 AI-default location names (33 tracked, −20% each) | | codexExemptions | (empty) | | found | (empty) | |
| 36.71% | AI-ism word frequency | Target: <2% AI-ism words (290 tracked) | | wordCount | 711 | | totalAiIsms | 9 | | found | | | highlights | | 0 | "dancing" | | 1 | "traced" | | 2 | "etched" | | 3 | "familiar" | | 4 | "pristine" | | 5 | "glint" | | 6 | "potential" | | 7 | "almost alive" | | 8 | "perfect" |
| |
| 100.00% | Cliché density | Target: ≤1 cliche(s) per 800-word window | | totalCliches | 0 | | maxInWindow | 0 | | found | (empty) | | highlights | (empty) | |
| 100.00% | Emotion telling (show vs. tell) | Target: ≤3% sentences with emotion telling | | emotionTells | 0 | | narrationSentences | 45 | | matches | (empty) | |
| 79.37% | Filter word density | Target: ≤3% sentences with filter/hedge words | | filterCount | 1 | | hedgeCount | 1 | | narrationSentences | 45 | | filterMatches | | | hedgeMatches | | |
| 100.00% | Gibberish response detection | Target: ≤1% gibberish-like sentences (hard fail if a sentence exceeds 800 words) | | analyzedSentences | 61 | | gibberishSentences | 0 | | adjustedGibberishSentences | 0 | | longSentenceCount | 0 | | runOnParagraphCount | 0 | | giantParagraphCount | 0 | | wordSaladCount | 0 | | repetitionLoopCount | 0 | | controlTokenCount | 0 | | maxSentenceWordsSeen | 30 | | ratio | 0 | | matches | (empty) | |
| 100.00% | Markdown formatting overuse | Target: ≤5% words in markdown formatting | | markdownSpans | 0 | | markdownWords | 0 | | totalWords | 715 | | ratio | 0 | | matches | (empty) | |
| 100.00% | Missing dialogue indicators (quotation marks) | Target: ≤10% speech attributions without quotation marks | | totalAttributions | 5 | | unquotedAttributions | 0 | | matches | (empty) | |
| 7.14% | Name drop frequency | Target: ≤1.0 per-name mentions per 100 words | | totalMentions | 28 | | wordCount | 455 | | uniqueNames | 11 | | maxNameDensity | 2.86 | | worstName | "Quinn" | | maxWindowNameDensity | 4.5 | | worstWindowName | "Quinn" | | discoveredNames | | Harlow | 1 | | Quinn | 13 | | Underground | 1 | | Sergeant | 1 | | Webb | 6 | | British | 1 | | Museum | 1 | | Eva | 1 | | Kowalski | 1 | | Full | 1 | | Morris | 1 |
| | persons | | 0 | "Harlow" | | 1 | "Quinn" | | 2 | "Sergeant" | | 3 | "Webb" | | 4 | "Eva" | | 5 | "Kowalski" | | 6 | "Morris" |
| | places | | | globalScore | 0.071 | | windowScore | 0.167 | |
| 100.00% | Narrator intent-glossing | Target: ≤2% narration sentences with intent-glossing patterns | | analyzedSentences | 33 | | glossingSentenceCount | 0 | | matches | (empty) | |
| 100.00% | "Not X but Y" pattern overuse | Target: ≤1 "not X but Y" per 1000 words | | totalMatches | 0 | | per1kWords | 0 | | wordCount | 715 | | matches | (empty) | |
| 100.00% | Overuse of "that" (subordinate clause padding) | Target: ≤2% sentences with "that" clauses | | thatCount | 0 | | totalSentences | 61 | | matches | (empty) | |
| 100.00% | Paragraph length variance | Target: CV ≥0.5 for paragraph word counts | | totalParagraphs | 28 | | mean | 25.54 | | std | 15.81 | | cv | 0.619 | | sampleLengths | | 0 | 39 | | 1 | 33 | | 2 | 30 | | 3 | 23 | | 4 | 9 | | 5 | 51 | | 6 | 6 | | 7 | 1 | | 8 | 19 | | 9 | 40 | | 10 | 7 | | 11 | 38 | | 12 | 46 | | 13 | 8 | | 14 | 43 | | 15 | 13 | | 16 | 29 | | 17 | 13 | | 18 | 49 | | 19 | 4 | | 20 | 46 | | 21 | 5 | | 22 | 27 | | 23 | 30 | | 24 | 31 | | 25 | 17 | | 26 | 48 | | 27 | 10 |
| |
| 89.67% | Passive voice overuse | Target: ≤2% passive sentences | | passiveCount | 2 | | totalSentences | 45 | | matches | | 0 | "was gone" | | 1 | "been sealed" |
| |
| 100.00% | Past progressive (was/were + -ing) overuse | Target: ≤2% past progressive verbs | | pastProgressiveCount | 0 | | totalVerbs | 76 | | matches | (empty) | |
| 0.00% | Em-dash & semicolon overuse | Target: ≤2% sentences with em-dashes/semicolons | | emDashCount | 4 | | semicolonCount | 0 | | flaggedSentences | 4 | | totalSentences | 61 | | ratio | 0.066 | | matches | | 0 | "The station had been sealed off for decades, yet someone had found their way in – and out." | | 1 | "The markings resembled the protective sigils she'd seen etched on that brass compass in evidence storage last month – the one they'd confiscated from that redheaded researcher at the British Museum." | | 2 | "The beam of her torch caught something in the tunnel – a glint of brass nestled between the rails." | | 3 | "They weren't random spatters – they formed a perfect circle, with precise geometric shapes inside." |
| |
| 100.00% | Purple prose (modifier overload) | Target: <4% adverbs, <2% -ly adverbs, no adj stacking | | wordCount | 451 | | adjectiveStacks | 0 | | stackExamples | (empty) | | adverbCount | 8 | | adverbRatio | 0.017738359201773836 | | lyAdverbCount | 4 | | lyAdverbRatio | 0.008869179600886918 | |
| 100.00% | Repeated phrase echo | Target: ≤20% sentences with echoes (window: 2) | | totalSentences | 61 | | echoCount | 0 | | echoWords | (empty) | |
| 100.00% | Sentence length variance | Target: CV ≥0.4 for sentence word counts | | totalSentences | 61 | | mean | 11.72 | | std | 6.35 | | cv | 0.542 | | sampleLengths | | 0 | 20 | | 1 | 19 | | 2 | 15 | | 3 | 18 | | 4 | 13 | | 5 | 17 | | 6 | 16 | | 7 | 7 | | 8 | 9 | | 9 | 4 | | 10 | 31 | | 11 | 16 | | 12 | 6 | | 13 | 1 | | 14 | 11 | | 15 | 8 | | 16 | 6 | | 17 | 22 | | 18 | 12 | | 19 | 7 | | 20 | 17 | | 21 | 21 | | 22 | 19 | | 23 | 9 | | 24 | 8 | | 25 | 10 | | 26 | 8 | | 27 | 4 | | 28 | 17 | | 29 | 12 | | 30 | 10 | | 31 | 7 | | 32 | 6 | | 33 | 15 | | 34 | 14 | | 35 | 13 | | 36 | 4 | | 37 | 23 | | 38 | 3 | | 39 | 19 | | 40 | 4 | | 41 | 11 | | 42 | 13 | | 43 | 22 | | 44 | 5 | | 45 | 9 | | 46 | 6 | | 47 | 12 | | 48 | 9 | | 49 | 15 |
| |
| 95.08% | Sentence opener variety | Target: ≥60% unique sentence openers | | consecutiveRepeats | 2 | | diversityRatio | 0.5901639344262295 | | totalSentences | 61 | | uniqueOpeners | 36 | |
| 0.00% | Adverb-first sentence starts | Target: ≥3% sentences starting with an adverb | | adverbCount | 0 | | totalSentences | 40 | | matches | (empty) | | ratio | 0 | |
| 100.00% | Pronoun-first sentence starts | Target: ≤30% sentences starting with a pronoun | | pronounCount | 9 | | totalSentences | 40 | | matches | | 0 | "Her hand instinctively went to" | | 1 | "She stood with military precision," | | 2 | "Its needle spun wildly, not" | | 3 | "She'd seen its twin in" | | 4 | "She pulled out her phone," | | 5 | "Its metal felt warm, almost" | | 6 | "She took one last look" | | 7 | "They weren't random spatters –" | | 8 | "She just stared into the" |
| | ratio | 0.225 | |
| 47.50% | Subject-first sentence starts | Target: ≤72% sentences starting with a subject | | subjectCount | 33 | | totalSentences | 40 | | matches | | 0 | "Detective Harlow Quinn knelt beside" | | 1 | "The body was gone, but" | | 2 | "The forensics team packed their" | | 3 | "The station had been sealed" | | 4 | "Detective Sergeant Webb shuffled through" | | 5 | "Quinn traced her finger along" | | 6 | "The markings resembled the protective" | | 7 | "Her hand instinctively went to" | | 8 | "She stood with military precision," | | 9 | "Webb's torch beam swept the" | | 10 | "Years of dust and debris" | | 11 | "Quinn walked the path, her" | | 12 | "The beam of her torch" | | 13 | "Quinn pulled on a latex" | | 14 | "A small compass, its casing" | | 15 | "Its needle spun wildly, not" | | 16 | "Quinn pocketed the compass." | | 17 | "She'd seen its twin in" | | 18 | "That researcher, Eva Kowalski, had" | | 19 | "Quinn stood at the platform's" |
| | ratio | 0.825 | |
| 0.00% | Subordinate conjunction sentence starts | Target: ≥2% sentences starting with a subordinating conjunction | | subConjCount | 0 | | totalSentences | 40 | | matches | (empty) | | ratio | 0 | |
| 100.00% | Technical jargon density | Target: ≤6% sentences with technical-jargon patterns | | analyzedSentences | 19 | | technicalSentenceCount | 1 | | matches | | 0 | "The body was gone, but dark stains remained on the crumbling tiles, forming patterns that defied natural blood spatter." |
| |
| 100.00% | Useless dialogue additions | Target: ≤5% dialogue tags with trailing filler fragments | | totalTags | 11 | | uselessAdditionCount | 0 | | matches | (empty) | |
| 38.89% | Dialogue tag variety (said vs. fancy) | Target: ≤10% fancy dialogue tags | | totalTags | 5 | | fancyCount | 3 | | fancyTags | | 0 | "Webb suggested (suggest)" | | 1 | "Webb continued (continue)" | | 2 | "Quinn ordered (order)" |
| | dialogueSentences | 27 | | tagDensity | 0.185 | | leniency | 0.37 | | rawRatio | 0.6 | | effectiveRatio | 0.222 | |