| 100.00% | Adverbs in dialogue tags | Target: ≤10% dialogue tags with adverbs | | totalTags | 17 | | adverbTagCount | 0 | | adverbTags | (empty) | | dialogueSentences | 38 | | tagDensity | 0.447 | | leniency | 0.895 | | rawRatio | 0 | | effectiveRatio | 0 | |
| 83.46% | AI-ism adverb frequency | Target: <2% AI-ism adverbs (58 tracked) | | wordCount | 1209 | | totalAiIsmAdverbs | 4 | | found | | | highlights | | 0 | "gently" | | 1 | "slowly" | | 2 | "carefully" | | 3 | "nervously" |
| |
| 60.00% | AI-ism character names | Target: 0 AI-default names (17 tracked, −20% each) | | codexExemptions | (empty) | | found | | |
| 100.00% | AI-ism location names | Target: 0 AI-default location names (33 tracked, −20% each) | | codexExemptions | (empty) | | found | (empty) | |
| 75.19% | AI-ism word frequency | Target: <2% AI-ism words (290 tracked) | | wordCount | 1209 | | totalAiIsms | 6 | | found | | | highlights | | 0 | "echoed" | | 1 | "processed" | | 2 | "perfect" | | 3 | "intricate" | | 4 | "etched" | | 5 | "determined" |
| |
| 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 | 72 | | matches | (empty) | |
| 100.00% | Filter word density | Target: ≤3% sentences with filter/hedge words | | filterCount | 0 | | hedgeCount | 2 | | narrationSentences | 72 | | filterMatches | (empty) | | hedgeMatches | | |
| 100.00% | Gibberish response detection | Target: ≤1% gibberish-like sentences (hard fail if a sentence exceeds 800 words) | | analyzedSentences | 93 | | gibberishSentences | 0 | | adjustedGibberishSentences | 0 | | longSentenceCount | 0 | | runOnParagraphCount | 0 | | giantParagraphCount | 0 | | wordSaladCount | 0 | | repetitionLoopCount | 0 | | controlTokenCount | 0 | | maxSentenceWordsSeen | 38 | | ratio | 0 | | matches | (empty) | |
| 100.00% | Markdown formatting overuse | Target: ≤5% words in markdown formatting | | markdownSpans | 0 | | markdownWords | 0 | | totalWords | 1216 | | ratio | 0 | | matches | (empty) | |
| 100.00% | Missing dialogue indicators (quotation marks) | Target: ≤10% speech attributions without quotation marks | | totalAttributions | 18 | | unquotedAttributions | 0 | | matches | (empty) | |
| 20.13% | Name drop frequency | Target: ≤1.0 per-name mentions per 100 words | | totalMentions | 48 | | wordCount | 847 | | uniqueNames | 15 | | maxNameDensity | 2.6 | | worstName | "Quinn" | | maxWindowNameDensity | 4 | | worstWindowName | "Quinn" | | discoveredNames | | Detective | 1 | | Harlow | 1 | | Quinn | 22 | | Tube | 1 | | Camden | 2 | | Chen | 7 | | Oyster | 1 | | Shoreditch | 2 | | Morris | 2 | | Eva | 3 | | Kowalski | 2 | | Sarah | 1 | | Morrison | 1 | | Covent | 1 | | Garden | 1 |
| | persons | | 0 | "Harlow" | | 1 | "Quinn" | | 2 | "Chen" | | 3 | "Morris" | | 4 | "Eva" | | 5 | "Kowalski" | | 6 | "Sarah" | | 7 | "Morrison" |
| | places | | 0 | "Shoreditch" | | 1 | "Covent" | | 2 | "Garden" |
| | globalScore | 0.201 | | windowScore | 0.333 | |
| 62.28% | Narrator intent-glossing | Target: ≤2% narration sentences with intent-glossing patterns | | analyzedSentences | 57 | | glossingSentenceCount | 2 | | matches | | 0 | "quite place" | | 1 | "as if checking for eavesdroppers among the uniformed officers" |
| |
| 100.00% | "Not X but Y" pattern overuse | Target: ≤1 "not X but Y" per 1000 words | | totalMatches | 0 | | per1kWords | 0 | | wordCount | 1216 | | matches | (empty) | |
| 100.00% | Overuse of "that" (subordinate clause padding) | Target: ≤2% sentences with "that" clauses | | thatCount | 0 | | totalSentences | 93 | | matches | (empty) | |
| 88.97% | Paragraph length variance | Target: CV ≥0.5 for paragraph word counts | | totalParagraphs | 39 | | mean | 31.18 | | std | 14.39 | | cv | 0.461 | | sampleLengths | | 0 | 50 | | 1 | 13 | | 2 | 44 | | 3 | 18 | | 4 | 30 | | 5 | 44 | | 6 | 20 | | 7 | 54 | | 8 | 52 | | 9 | 24 | | 10 | 30 | | 11 | 52 | | 12 | 17 | | 13 | 46 | | 14 | 34 | | 15 | 22 | | 16 | 36 | | 17 | 23 | | 18 | 5 | | 19 | 37 | | 20 | 38 | | 21 | 8 | | 22 | 32 | | 23 | 53 | | 24 | 36 | | 25 | 7 | | 26 | 19 | | 27 | 38 | | 28 | 25 | | 29 | 55 | | 30 | 25 | | 31 | 46 | | 32 | 7 | | 33 | 29 | | 34 | 34 | | 35 | 20 | | 36 | 42 | | 37 | 12 | | 38 | 39 |
| |
| 85.77% | Passive voice overuse | Target: ≤2% passive sentences | | passiveCount | 4 | | totalSentences | 72 | | matches | | 0 | "were folded" | | 1 | "been drawn" | | 2 | "been drawn" | | 3 | "was disturbed" |
| |
| 100.00% | Past progressive (was/were + -ing) overuse | Target: ≤2% past progressive verbs | | pastProgressiveCount | 1 | | totalVerbs | 144 | | matches | | |
| 0.00% | Em-dash & semicolon overuse | Target: ≤2% sentences with em-dashes/semicolons | | emDashCount | 9 | | semicolonCount | 0 | | flaggedSentences | 8 | | totalSentences | 93 | | ratio | 0.086 | | matches | | 0 | "The stench hit Detective Harlow Quinn before she even reached the bottom of the maintenance ladder — a sweet, cloying rot mixed with something chemical that made her eyes water." | | 1 | "The victim — a woman in her twenties — sat propped against the tunnel wall in an almost meditative pose." | | 2 | "She noted the victim's clothing — an expensive wool coat, designer boots." | | 3 | "She checked the coat pockets — empty except for an Oyster card and a receipt from a coffee shop in Shoreditch dated two days ago." | | 4 | "Not just any compass — the brass casing was green with verdigris, and when she lifted it carefully with an evidence bag, she could see strange markings etched into its face." | | 5 | "He was right — there was a pattern within the chaos, something that might be linguistic rather than purely decorative." | | 6 | "Eva Kowalski tucked a strand of hair behind her left ear — a nervous gesture Quinn catalogued automatically." | | 7 | "Quinn pocketed her phone and followed Eva Kowalski into the café, unaware that she was about to step into a different kind of darkness entirely — one that had been waiting for her since the night DS Morris died." |
| |
| 100.00% | Purple prose (modifier overload) | Target: <4% adverbs, <2% -ly adverbs, no adj stacking | | wordCount | 842 | | adjectiveStacks | 0 | | stackExamples | (empty) | | adverbCount | 33 | | adverbRatio | 0.039192399049881234 | | lyAdverbCount | 11 | | lyAdverbRatio | 0.013064133016627079 | |
| 100.00% | Repeated phrase echo | Target: ≤20% sentences with echoes (window: 2) | | totalSentences | 93 | | echoCount | 0 | | echoWords | (empty) | |
| 100.00% | Sentence length variance | Target: CV ≥0.4 for sentence word counts | | totalSentences | 93 | | mean | 13.08 | | std | 7.53 | | cv | 0.576 | | sampleLengths | | 0 | 30 | | 1 | 20 | | 2 | 11 | | 3 | 2 | | 4 | 15 | | 5 | 10 | | 6 | 19 | | 7 | 18 | | 8 | 25 | | 9 | 5 | | 10 | 7 | | 11 | 15 | | 12 | 22 | | 13 | 6 | | 14 | 14 | | 15 | 7 | | 16 | 20 | | 17 | 16 | | 18 | 11 | | 19 | 18 | | 20 | 16 | | 21 | 18 | | 22 | 9 | | 23 | 12 | | 24 | 3 | | 25 | 12 | | 26 | 18 | | 27 | 13 | | 28 | 10 | | 29 | 4 | | 30 | 25 | | 31 | 12 | | 32 | 5 | | 33 | 10 | | 34 | 18 | | 35 | 18 | | 36 | 5 | | 37 | 22 | | 38 | 7 | | 39 | 22 | | 40 | 12 | | 41 | 18 | | 42 | 6 | | 43 | 5 | | 44 | 18 | | 45 | 3 | | 46 | 2 | | 47 | 16 | | 48 | 11 | | 49 | 10 |
| |
| 93.19% | Sentence opener variety | Target: ≥60% unique sentence openers | | consecutiveRepeats | 0 | | diversityRatio | 0.5591397849462365 | | totalSentences | 93 | | uniqueOpeners | 52 | |
| 100.00% | Adverb-first sentence starts | Target: ≥3% sentences starting with an adverb | | adverbCount | 2 | | totalSentences | 65 | | matches | | 0 | "Of course someone had leaked" | | 1 | "Just another young woman who'd" |
| | ratio | 0.031 | |
| 100.00% | Pronoun-first sentence starts | Target: ≤30% sentences starting with a pronoun | | pronounCount | 14 | | totalSentences | 65 | | matches | | 0 | "Her boots splashed into ankle-deep" | | 1 | "They sloshed forward through the" | | 2 | "Her hands were folded neatly" | | 3 | "She noted the victim's clothing" | | 4 | "She checked the coat pockets" | | 5 | "She pulled out her own" | | 6 | "She turned to the forensics" | | 7 | "She waded over and knelt," | | 8 | "she told the technician" | | 9 | "he said, pointing to a" | | 10 | "He was right — there" | | 11 | "She thought of DS Morris," | | 12 | "She started back toward the" | | 13 | "Her instincts, honed over eighteen" |
| | ratio | 0.215 | |
| 90.77% | Subject-first sentence starts | Target: ≤72% sentences starting with a subject | | subjectCount | 48 | | totalSentences | 65 | | matches | | 0 | "The stench hit Detective Harlow" | | 1 | "Her boots splashed into ankle-deep" | | 2 | "DS Chen called down" | | 3 | "Quinn swept her torch across" | | 4 | "The curved walls glistened with" | | 5 | "Quinn asked as Chen descended" | | 6 | "Chen's voice echoed off the" | | 7 | "They sloshed forward through the" | | 8 | "Quinn kept her breathing shallow," | | 9 | "the lead technician nodded" | | 10 | "Quinn stepped closer, and her" | | 11 | "The victim — a woman" | | 12 | "Her hands were folded neatly" | | 13 | "A perfect circle had been" | | 14 | "The symbols seemed to shimmer" | | 15 | "Quinn asked, crouching beside the" | | 16 | "She noted the victim's clothing" | | 17 | "the technician said" | | 18 | "Quinn pulled on latex gloves" | | 19 | "Rigor had set in, but" |
| | ratio | 0.738 | |
| 0.00% | Subordinate conjunction sentence starts | Target: ≥2% sentences starting with a subordinating conjunction | | subConjCount | 0 | | totalSentences | 65 | | matches | (empty) | | ratio | 0 | |
| 100.00% | Technical jargon density | Target: ≤6% sentences with technical-jargon patterns | | analyzedSentences | 38 | | technicalSentenceCount | 1 | | matches | | 0 | "He was right — there was a pattern within the chaos, something that might be linguistic rather than purely decorative." |
| |
| 100.00% | Useless dialogue additions | Target: ≤5% dialogue tags with trailing filler fragments | | totalTags | 17 | | uselessAdditionCount | 0 | | matches | (empty) | |
| 100.00% | Dialogue tag variety (said vs. fancy) | Target: ≤10% fancy dialogue tags | | totalTags | 10 | | fancyCount | 1 | | fancyTags | | 0 | "Chen suggested (suggest)" |
| | dialogueSentences | 38 | | tagDensity | 0.263 | | leniency | 0.526 | | rawRatio | 0.1 | | effectiveRatio | 0.053 | |