| 100.00% | Adverbs in dialogue tags | Target: ≤10% dialogue tags with adverbs | | totalTags | 10 | | adverbTagCount | 0 | | adverbTags | (empty) | | dialogueSentences | 29 | | tagDensity | 0.345 | | leniency | 0.69 | | rawRatio | 0 | | effectiveRatio | 0 | |
| 96.84% | AI-ism adverb frequency | Target: <2% AI-ism adverbs (58 tracked) | | wordCount | 1580 | | 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) | |
| 43.04% | AI-ism word frequency | Target: <2% AI-ism words (290 tracked) | | wordCount | 1580 | | totalAiIsms | 18 | | found | | | highlights | | 0 | "pulse" | | 1 | "fleeting" | | 2 | "scanning" | | 3 | "pulsed" | | 4 | "footsteps" | | 5 | "perfect" | | 6 | "flicker" | | 7 | "trembled" | | 8 | "tinged" | | 9 | "flicked" | | 10 | "electric" | | 11 | "glistening" | | 12 | "beacon" | | 13 | "weight" |
| |
| 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 | 140 | | matches | (empty) | |
| 100.00% | Filter word density | Target: ≤3% sentences with filter/hedge words | | filterCount | 2 | | hedgeCount | 0 | | narrationSentences | 140 | | filterMatches | | | hedgeMatches | (empty) | |
| 100.00% | Gibberish response detection | Target: ≤1% gibberish-like sentences (hard fail if a sentence exceeds 800 words) | | analyzedSentences | 160 | | gibberishSentences | 0 | | adjustedGibberishSentences | 0 | | longSentenceCount | 0 | | runOnParagraphCount | 0 | | giantParagraphCount | 0 | | wordSaladCount | 0 | | repetitionLoopCount | 0 | | controlTokenCount | 0 | | maxSentenceWordsSeen | 35 | | ratio | 0 | | matches | (empty) | |
| 100.00% | Markdown formatting overuse | Target: ≤5% words in markdown formatting | | markdownSpans | 23 | | markdownWords | 53 | | totalWords | 1564 | | ratio | 0.034 | | matches | | 0 | "The Raven’s Nest" | | 1 | "target" | | 2 | "suspect" | | 3 | "not again" | | 4 | "OUT OF ORDER" | | 5 | "The Raven’s Nest" | | 6 | "wrong" | | 7 | "Back-up needed, chasing a bloke with a magic bone into a hole in the ground" | | 8 | "“Quinn, this isn’t—”" | | 9 | "intentional" | | 10 | "The Veil Market" | | 11 | "sure" | | 12 | "wrong" | | 13 | "jurisdiction" | | 14 | "detective" | | 15 | "detective" | | 16 | "“Some doors shouldn’t be opened.”" | | 17 | "burned" | | 18 | "idiot" | | 19 | "Now." | | 20 | "Now." | | 21 | "no idea" | | 22 | "beacon" |
| |
| 100.00% | Missing dialogue indicators (quotation marks) | Target: ≤10% speech attributions without quotation marks | | totalAttributions | 15 | | unquotedAttributions | 0 | | matches | (empty) | |
| 79.92% | Name drop frequency | Target: ≤1.0 per-name mentions per 100 words | | totalMentions | 55 | | wordCount | 1427 | | uniqueNames | 16 | | maxNameDensity | 1.4 | | worstName | "Harlow" | | maxWindowNameDensity | 2.5 | | worstWindowName | "Harlow" | | discoveredNames | | Harlow | 20 | | Quinn | 1 | | Soho | 2 | | Raven | 2 | | London | 1 | | Morris | 5 | | Wardour | 1 | | Street | 1 | | Saturday-night | 1 | | Veil | 1 | | Tommy | 11 | | Herrera | 2 | | Spanish | 2 | | Saint | 1 | | Christopher | 1 | | Three | 3 |
| | persons | | 0 | "Harlow" | | 1 | "Quinn" | | 2 | "Raven" | | 3 | "Morris" | | 4 | "Tommy" | | 5 | "Herrera" | | 6 | "Saint" | | 7 | "Christopher" |
| | places | | 0 | "Soho" | | 1 | "London" | | 2 | "Wardour" | | 3 | "Street" | | 4 | "Spanish" |
| | globalScore | 0.799 | | windowScore | 0.833 | |
| 92.53% | Narrator intent-glossing | Target: ≤2% narration sentences with intent-glossing patterns | | analyzedSentences | 87 | | glossingSentenceCount | 2 | | matches | | 0 | "looked like preserved eyes" | | 1 | "seemed longer on the way up" |
| |
| 100.00% | "Not X but Y" pattern overuse | Target: ≤1 "not X but Y" per 1000 words | | totalMatches | 0 | | per1kWords | 0 | | wordCount | 1564 | | matches | (empty) | |
| 100.00% | Overuse of "that" (subordinate clause padding) | Target: ≤2% sentences with "that" clauses | | thatCount | 0 | | totalSentences | 160 | | matches | (empty) | |
| 100.00% | Paragraph length variance | Target: CV ≥0.5 for paragraph word counts | | totalParagraphs | 63 | | mean | 24.83 | | std | 25.05 | | cv | 1.009 | | sampleLengths | | 0 | 90 | | 1 | 55 | | 2 | 18 | | 3 | 87 | | 4 | 71 | | 5 | 5 | | 6 | 77 | | 7 | 38 | | 8 | 5 | | 9 | 67 | | 10 | 37 | | 11 | 4 | | 12 | 2 | | 13 | 66 | | 14 | 11 | | 15 | 66 | | 16 | 68 | | 17 | 6 | | 18 | 4 | | 19 | 88 | | 20 | 27 | | 21 | 27 | | 22 | 7 | | 23 | 6 | | 24 | 2 | | 25 | 38 | | 26 | 2 | | 27 | 39 | | 28 | 16 | | 29 | 15 | | 30 | 7 | | 31 | 22 | | 32 | 4 | | 33 | 9 | | 34 | 48 | | 35 | 9 | | 36 | 12 | | 37 | 47 | | 38 | 6 | | 39 | 6 | | 40 | 6 | | 41 | 2 | | 42 | 43 | | 43 | 9 | | 44 | 18 | | 45 | 7 | | 46 | 15 | | 47 | 2 | | 48 | 1 | | 49 | 40 |
| |
| 90.23% | Passive voice overuse | Target: ≤2% passive sentences | | passiveCount | 6 | | totalSentences | 140 | | matches | | 0 | "was gone" | | 1 | "was gone" | | 2 | "been chased" | | 3 | "was gone" | | 4 | "was gone" | | 5 | "was written" |
| |
| 100.00% | Past progressive (was/were + -ing) overuse | Target: ≤2% past progressive verbs | | pastProgressiveCount | 1 | | totalVerbs | 265 | | matches | | |
| 0.00% | Em-dash & semicolon overuse | Target: ≤2% sentences with em-dashes/semicolons | | emDashCount | 18 | | semicolonCount | 0 | | flaggedSentences | 12 | | totalSentences | 160 | | ratio | 0.075 | | matches | | 0 | "The only thing that mattered was the figure ahead—tall, lean, moving with the unnatural grace of someone who knew these streets better than their own pulse." | | 1 | "Rain hammered the cobblestones, turning the alley into a gauntlet of reflections—her own face, distorted and fleeting, flashed in a puddle as she sprinted past." | | 2 | "The suspect—*target*, her brain corrected, because *suspect* implied rules, and rules didn’t apply here—ducked left into a narrow passage, his dark coat flaring like wings." | | 3 | "She palmed her warrant card as she ran, flashing it at a bouncer whose face twisted in recognition—*not again*—before he stepped aside." | | 4 | "Not footsteps—too light, too *wrong*." | | 5 | "The torchlight caught movement—just a flicker, there and gone." | | 6 | "The air was thicker here, heavy with the scent of burnt herbs and something coppery—blood, but old." | | 7 | "And the way the target had moved—too fast, too *sure*—was real." | | 8 | "Handing the bone token to a figure draped in a hooded cloak, their fingers—too long, too *wrong*—closing around it." | | 9 | "The voice was smooth, accented—Spanish, maybe—and entirely too calm for a man who’d just been chased through half of Soho." | | 10 | "Harlow’s watch—Morris’ watch—*burned* against her wrist." | | 11 | "The liquid inside sloshed, thick and dark, and the label—if it was a label—was written in a script that made her eyes ache." |
| |
| 94.00% | Purple prose (modifier overload) | Target: <4% adverbs, <2% -ly adverbs, no adj stacking | | wordCount | 457 | | adjectiveStacks | 1 | | stackExamples | | 0 | "ahead—tall, lean, moving" |
| | adverbCount | 14 | | adverbRatio | 0.030634573304157548 | | lyAdverbCount | 6 | | lyAdverbRatio | 0.01312910284463895 | |
| 100.00% | Repeated phrase echo | Target: ≤20% sentences with echoes (window: 2) | | totalSentences | 160 | | echoCount | 0 | | echoWords | (empty) | |
| 100.00% | Sentence length variance | Target: CV ≥0.4 for sentence word counts | | totalSentences | 160 | | mean | 9.78 | | std | 7.64 | | cv | 0.782 | | sampleLengths | | 0 | 26 | | 1 | 24 | | 2 | 14 | | 3 | 26 | | 4 | 3 | | 5 | 17 | | 6 | 35 | | 7 | 11 | | 8 | 3 | | 9 | 4 | | 10 | 25 | | 11 | 25 | | 12 | 11 | | 13 | 26 | | 14 | 25 | | 15 | 20 | | 16 | 4 | | 17 | 22 | | 18 | 5 | | 19 | 8 | | 20 | 12 | | 21 | 18 | | 22 | 20 | | 23 | 19 | | 24 | 2 | | 25 | 5 | | 26 | 6 | | 27 | 16 | | 28 | 3 | | 29 | 1 | | 30 | 1 | | 31 | 4 | | 32 | 5 | | 33 | 17 | | 34 | 19 | | 35 | 11 | | 36 | 15 | | 37 | 1 | | 38 | 4 | | 39 | 9 | | 40 | 12 | | 41 | 16 | | 42 | 1 | | 43 | 3 | | 44 | 2 | | 45 | 7 | | 46 | 8 | | 47 | 8 | | 48 | 25 | | 49 | 2 |
| |
| 51.46% | Sentence opener variety | Target: ≥60% unique sentence openers | | consecutiveRepeats | 11 | | diversityRatio | 0.35 | | totalSentences | 160 | | uniqueOpeners | 56 | |
| 100.00% | Adverb-first sentence starts | Target: ≥3% sentences starting with an adverb | | adverbCount | 6 | | totalSentences | 121 | | matches | | 0 | "Of course he didn’t." | | 1 | "Then the target was gone." | | 2 | "Just his watch, still ticking," | | 3 | "Then she descended." | | 4 | "Of course he knew." | | 5 | "Then the figure raised a" |
| | ratio | 0.05 | |
| 100.00% | Pronoun-first sentence starts | Target: ≤30% sentences starting with a pronoun | | pronounCount | 24 | | totalSentences | 121 | | matches | | 0 | "She hit the alleyway in" | | 1 | "He didn’t stop." | | 2 | "She palmed her warrant card" | | 3 | "She knew because she’d mapped" | | 4 | "She drew her torch, the" | | 5 | "Her fingers twitched toward her" | | 6 | "He’d turned to her, his" | | 7 | "Her skin prickled." | | 8 | "She’d dismissed it as junkie" | | 9 | "She was out of her" | | 10 | "She reached for her cuffs." | | 11 | "His dark curls were damp," | | 12 | "He’d been the one to" | | 13 | "She didn’t look back." | | 14 | "She didn’t scream." | | 15 | "She didn’t run." | | 16 | "Her shoulder collided with Tommy" | | 17 | "His fingers dug into her" | | 18 | "She let him pull her" | | 19 | "His curse was fluent, creative," |
| | ratio | 0.198 | |
| 88.10% | Subject-first sentence starts | Target: ≤72% sentences starting with a subject | | subjectCount | 90 | | totalSentences | 121 | | matches | | 0 | "The pavement cracked beneath Harlow" | | 1 | "She hit the alleyway in" | | 2 | "The only thing that mattered" | | 3 | "A bone token." | | 4 | "That’s what she’d seen in" | | 5 | "The words tore from her" | | 6 | "He didn’t stop." | | 7 | "The suspect—*target*, her brain corrected," | | 8 | "Harlow followed, her breath steady" | | 9 | "The passage spat them out" | | 10 | "A woman in a sequined" | | 11 | "Harlow didn’t break stride." | | 12 | "She palmed her warrant card" | | 13 | "Harlow skidded to a halt," | | 14 | "She knew because she’d mapped" | | 15 | "A scrape, like metal on" | | 16 | "Harlow turned just as a" | | 17 | "The car park." | | 18 | "She drew her torch, the" | | 19 | "The air smelled of petrol" |
| | ratio | 0.744 | |
| 0.00% | Subordinate conjunction sentence starts | Target: ≥2% sentences starting with a subordinating conjunction | | subConjCount | 0 | | totalSentences | 121 | | matches | (empty) | | ratio | 0 | |
| 48.52% | Technical jargon density | Target: ≤6% sentences with technical-jargon patterns | | analyzedSentences | 53 | | technicalSentenceCount | 7 | | matches | | 0 | "The only thing that mattered was the figure ahead—tall, lean, moving with the unnatural grace of someone who knew these streets better than their own pulse." | | 1 | "Small, carved, the kind of thing that didn’t belong in a detective’s evidence bag but ended up there anyway, because London had layers, and Harlow had long sinc…" | | 2 | "A woman in a sequined dress shrieked as the target shouldered past her, sending her clutch spinning into the gutter." | | 3 | "The air smelled of petrol and damp, the kind of scent that clung to the back of your throat." | | 4 | "The beam of her torch trembled as she swept it left, right, catching glimpses of graffiti that wasn’t graffiti: symbols carved into the stone, their edges too p…" | | 5 | "Light bled from lanterns hung on rusted hooks, their flames blue-tinged, casting long shadows over stalls that shouldn’t have fit in the space." | | 6 | "The air hummed with the low thrum of bartering, the clink of coins that weren’t coins, the scent of spices that made her head swim." |
| |
| 75.00% | Useless dialogue additions | Target: ≤5% dialogue tags with trailing filler fragments | | totalTags | 10 | | uselessAdditionCount | 1 | | matches | | 0 | "The Raven’s Nest* pulsed like, but Silas’ bar was a dead end" |
| |
| 81.03% | Dialogue tag variety (said vs. fancy) | Target: ≤10% fancy dialogue tags | | totalTags | 3 | | fancyCount | 2 | | fancyTags | | 0 | "The one who’d warned her (had warn)" | | 1 | "she demanded (demand)" |
| | dialogueSentences | 29 | | tagDensity | 0.103 | | leniency | 0.207 | | rawRatio | 0.667 | | effectiveRatio | 0.138 | |