| 100.00% | Adverbs in dialogue tags | Target: ≤10% dialogue tags with adverbs | | totalTags | 15 | | adverbTagCount | 1 | | adverbTags | | 0 | "Davies gestured vaguely [vaguely]" |
| | dialogueSentences | 46 | | tagDensity | 0.326 | | leniency | 0.652 | | rawRatio | 0.067 | | effectiveRatio | 0.043 | |
| 70.16% | AI-ism adverb frequency | Target: <2% AI-ism adverbs (58 tracked) | | wordCount | 1508 | | totalAiIsmAdverbs | 9 | | found | | 0 | | | 1 | | | 2 | | | 3 | | adverb | "deliberately" | | count | 2 |
|
| | highlights | | 0 | "slightly" | | 1 | "carefully" | | 2 | "very" | | 3 | "deliberately" |
| |
| 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) | |
| 73.47% | AI-ism word frequency | Target: <2% AI-ism words (290 tracked) | | wordCount | 1508 | | totalAiIsms | 8 | | found | | | highlights | | 0 | "standard" | | 1 | "tension" | | 2 | "glinting" | | 3 | "resolved" | | 4 | "etched" | | 5 | "echoed" | | 6 | "constructed" |
| |
| 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 | 116 | | matches | (empty) | |
| 93.60% | Filter word density | Target: ≤3% sentences with filter/hedge words | | filterCount | 2 | | hedgeCount | 2 | | narrationSentences | 116 | | filterMatches | | | hedgeMatches | | |
| 100.00% | Gibberish response detection | Target: ≤1% gibberish-like sentences (hard fail if a sentence exceeds 800 words) | | analyzedSentences | 147 | | gibberishSentences | 0 | | adjustedGibberishSentences | 0 | | longSentenceCount | 0 | | runOnParagraphCount | 0 | | giantParagraphCount | 0 | | wordSaladCount | 0 | | repetitionLoopCount | 0 | | controlTokenCount | 0 | | maxSentenceWordsSeen | 54 | | ratio | 0 | | matches | (empty) | |
| 100.00% | Markdown formatting overuse | Target: ≤5% words in markdown formatting | | markdownSpans | 1 | | markdownWords | 10 | | totalWords | 1506 | | ratio | 0.007 | | matches | | 0 | "Who is trying to open the Veil from our side?" |
| |
| 100.00% | Missing dialogue indicators (quotation marks) | Target: ≤10% speech attributions without quotation marks | | totalAttributions | 13 | | unquotedAttributions | 0 | | matches | (empty) | |
| 66.67% | Name drop frequency | Target: ≤1.0 per-name mentions per 100 words | | totalMentions | 49 | | wordCount | 1124 | | uniqueNames | 16 | | maxNameDensity | 1.6 | | worstName | "Quinn" | | maxWindowNameDensity | 3 | | worstWindowName | "Quinn" | | discoveredNames | | Chapter | 1 | | Geometry | 1 | | Wrongness | 1 | | Harlow | 1 | | Quinn | 18 | | Camden | 1 | | Metropolitan | 2 | | Police | 2 | | Davies | 13 | | High-quality | 1 | | Face | 1 | | Coleman | 1 | | Morris | 2 | | Veil | 2 | | Compass | 1 | | London | 1 |
| | persons | | 0 | "Harlow" | | 1 | "Quinn" | | 2 | "Police" | | 3 | "Davies" | | 4 | "Face" | | 5 | "Coleman" | | 6 | "Morris" |
| | places | | | globalScore | 0.699 | | windowScore | 0.667 | |
| 83.33% | Narrator intent-glossing | Target: ≤2% narration sentences with intent-glossing patterns | | analyzedSentences | 75 | | glossingSentenceCount | 2 | | matches | | 0 | "seemed wrong" | | 1 | "felt like" |
| |
| 100.00% | "Not X but Y" pattern overuse | Target: ≤1 "not X but Y" per 1000 words | | totalMatches | 0 | | per1kWords | 0 | | wordCount | 1506 | | matches | (empty) | |
| 100.00% | Overuse of "that" (subordinate clause padding) | Target: ≤2% sentences with "that" clauses | | thatCount | 1 | | totalSentences | 147 | | matches | | |
| 100.00% | Paragraph length variance | Target: CV ≥0.5 for paragraph word counts | | totalParagraphs | 59 | | mean | 25.53 | | std | 22.72 | | cv | 0.89 | | sampleLengths | | 0 | 7 | | 1 | 7 | | 2 | 59 | | 3 | 12 | | 4 | 46 | | 5 | 48 | | 6 | 14 | | 7 | 41 | | 8 | 49 | | 9 | 21 | | 10 | 4 | | 11 | 43 | | 12 | 2 | | 13 | 32 | | 14 | 28 | | 15 | 41 | | 16 | 6 | | 17 | 51 | | 18 | 56 | | 19 | 11 | | 20 | 5 | | 21 | 5 | | 22 | 7 | | 23 | 80 | | 24 | 6 | | 25 | 7 | | 26 | 11 | | 27 | 7 | | 28 | 14 | | 29 | 3 | | 30 | 11 | | 31 | 48 | | 32 | 6 | | 33 | 2 | | 34 | 5 | | 35 | 2 | | 36 | 7 | | 37 | 19 | | 38 | 11 | | 39 | 52 | | 40 | 2 | | 41 | 31 | | 42 | 65 | | 43 | 38 | | 44 | 13 | | 45 | 42 | | 46 | 9 | | 47 | 23 | | 48 | 17 | | 49 | 3 |
| |
| 96.19% | Passive voice overuse | Target: ≤2% passive sentences | | passiveCount | 3 | | totalSentences | 116 | | matches | | 0 | "was staged" | | 1 | "been struck" | | 2 | "was etched" |
| |
| 100.00% | Past progressive (was/were + -ing) overuse | Target: ≤2% past progressive verbs | | pastProgressiveCount | 1 | | totalVerbs | 191 | | matches | | |
| 100.00% | Em-dash & semicolon overuse | Target: ≤2% sentences with em-dashes/semicolons | | emDashCount | 0 | | semicolonCount | 0 | | flaggedSentences | 0 | | totalSentences | 147 | | ratio | 0 | | matches | (empty) | |
| 100.00% | Purple prose (modifier overload) | Target: <4% adverbs, <2% -ly adverbs, no adj stacking | | wordCount | 1126 | | adjectiveStacks | 0 | | stackExamples | (empty) | | adverbCount | 42 | | adverbRatio | 0.037300177619893425 | | lyAdverbCount | 14 | | lyAdverbRatio | 0.012433392539964476 | |
| 100.00% | Repeated phrase echo | Target: ≤20% sentences with echoes (window: 2) | | totalSentences | 147 | | echoCount | 0 | | echoWords | (empty) | |
| 100.00% | Sentence length variance | Target: CV ≥0.4 for sentence word counts | | totalSentences | 147 | | mean | 10.24 | | std | 8.33 | | cv | 0.813 | | sampleLengths | | 0 | 14 | | 1 | 29 | | 2 | 9 | | 3 | 5 | | 4 | 16 | | 5 | 9 | | 6 | 3 | | 7 | 32 | | 8 | 14 | | 9 | 5 | | 10 | 18 | | 11 | 4 | | 12 | 3 | | 13 | 18 | | 14 | 14 | | 15 | 33 | | 16 | 8 | | 17 | 9 | | 18 | 19 | | 19 | 21 | | 20 | 11 | | 21 | 10 | | 22 | 4 | | 23 | 18 | | 24 | 2 | | 25 | 2 | | 26 | 10 | | 27 | 5 | | 28 | 6 | | 29 | 2 | | 30 | 6 | | 31 | 3 | | 32 | 5 | | 33 | 8 | | 34 | 7 | | 35 | 3 | | 36 | 3 | | 37 | 25 | | 38 | 8 | | 39 | 3 | | 40 | 2 | | 41 | 3 | | 42 | 25 | | 43 | 6 | | 44 | 7 | | 45 | 8 | | 46 | 13 | | 47 | 2 | | 48 | 3 | | 49 | 10 |
| |
| 56.01% | Sentence opener variety | Target: ≥60% unique sentence openers | | consecutiveRepeats | 16 | | diversityRatio | 0.4013605442176871 | | totalSentences | 147 | | uniqueOpeners | 59 | |
| 64.10% | Adverb-first sentence starts | Target: ≥3% sentences starting with an adverb | | adverbCount | 2 | | totalSentences | 104 | | matches | | 0 | "More expensive equipment." | | 1 | "Then she photographed the corpse" |
| | ratio | 0.019 | |
| 100.00% | Pronoun-first sentence starts | Target: ≤30% sentences starting with a pronoun | | pronounCount | 22 | | totalSentences | 104 | | matches | | 0 | "She'd been a Metropolitan Police" | | 1 | "She noticed things." | | 2 | "She pulled on latex gloves," | | 3 | "She repeated the word, testing" | | 4 | "She touched nothing, just observed." | | 5 | "She moved to the next" | | 6 | "She'd learned to trust that" | | 7 | "She pulled out her phone" | | 8 | "She shifted position, moving her" | | 9 | "They painted a story if" | | 10 | "She pushed the thought away." | | 11 | "She pulled out a compact" | | 12 | "She descended the metal ladder" | | 13 | "She turned, training her torch" | | 14 | "She'd spent three years learning" | | 15 | "She knew what this was." | | 16 | "She bagged the compass without" | | 17 | "Her mind was already three" | | 18 | "She climbed back up to" | | 19 | "She glanced at her watch," |
| | ratio | 0.212 | |
| 51.35% | Subject-first sentence starts | Target: ≤72% sentences starting with a subject | | subjectCount | 85 | | totalSentences | 104 | | matches | | 0 | "# Chapter 2: The Geometry" | | 1 | "Detective Harlow Quinn crouched beside" | | 2 | "The outline curved slightly inward" | | 3 | "The actual body curved outward." | | 4 | "Someone had drawn it from" | | 5 | "She'd been a Metropolitan Police" | | 6 | "She noticed things." | | 7 | "Sergeant Davies said from behind" | | 8 | "Quinn stood, her knees protesting." | | 9 | "The station smelled of rust," | | 10 | "Human waste, she realised." | | 11 | "Someone lived here." | | 12 | "She pulled on latex gloves," | | 13 | "Davies gestured vaguely at the" | | 14 | "Quinn began walking the perimeter" | | 15 | "The work lights cast long" | | 16 | "The tiled walls bore the" | | 17 | "She repeated the word, testing" | | 18 | "Quinn crouched beside a sleeping" | | 19 | "She touched nothing, just observed." |
| | ratio | 0.817 | |
| 100.00% | Subordinate conjunction sentence starts | Target: ≥2% sentences starting with a subordinating conjunction | | subConjCount | 3 | | totalSentences | 104 | | matches | | 0 | "Even with the work lights," | | 1 | "By the time Davies returned" | | 2 | "Because that's what this felt" |
| | ratio | 0.029 | |
| 63.49% | Technical jargon density | Target: ≤6% sentences with technical-jargon patterns | | analyzedSentences | 45 | | technicalSentenceCount | 5 | | matches | | 0 | "The work lights cast long shadows, turning the abandoned tube station into something that belonged in a fever dream." | | 1 | "She'd learned to trust that feeling, the small tension at the hinge of her teeth that meant something had slipped sideways from the official narrative." | | 2 | "The spatter suggested he'd been struck from the side, his body rotating with the impact before collapsing." | | 3 | "The kind of thing that shouldn't exist outside of Davies's nightmare, let alone in a London tube station." | | 4 | "The mysterious structural assessment that had brought a worker into the station on a weekend." |
| |
| 58.33% | Useless dialogue additions | Target: ≤5% dialogue tags with trailing filler fragments | | totalTags | 15 | | uselessAdditionCount | 2 | | matches | | 0 | "She pulled, her movements precise and economical" | | 1 | "Quinn dropped, her knees absorbing the impact" |
| |
| 100.00% | Dialogue tag variety (said vs. fancy) | Target: ≤10% fancy dialogue tags | | totalTags | 7 | | fancyCount | 2 | | fancyTags | | 0 | "She repeated (repeat)" | | 1 | "Quinn repeated (repeat)" |
| | dialogueSentences | 46 | | tagDensity | 0.152 | | leniency | 0.304 | | rawRatio | 0.286 | | effectiveRatio | 0.087 | |