| 100.00% | Adverbs in dialogue tags | Target: ≤10% dialogue tags with adverbs | |
| 95.71% | AI-ism adverb frequency | Target: <2% AI-ism adverbs (58 tracked) | | wordCount | 2331 | | totalAiIsmAdverbs | 2 | | found | | | highlights | | |
| 100.00% | AI-ism character names | Target: 0 AI-default names (16 tracked, −20% each) | | codexExemptions | | | found | (empty) | |
| 100.00% | AI-ism location names | Target: 0 AI-default location names (33 tracked, −20% each) | | codexExemptions | (empty) | | found | (empty) | |
| 50.66% | AI-ism word frequency | Target: <2% AI-ism words (290 tracked) | | wordCount | 2331 | | totalAiIsms | 23 | | found | | | highlights | | 0 | "scanned" | | 1 | "weight" | | 2 | "flicked" | | 3 | "coded" | | 4 | "silence" | | 5 | "tracing" | | 6 | "reminder" | | 7 | "echoed" | | 8 | "familiar" | | 9 | "aligned" | | 10 | "echo" | | 11 | "unspoken" | | 12 | "traced" | | 13 | "beacon" | | 14 | "tension" | | 15 | "measured" |
| |
| 100.00% | Cliché density | Target: ≤1 cliche(s) per 800-word window | | totalCliches | 1 | | maxInWindow | 1 | | found | | 0 | | label | "eyes widened/narrowed" | | count | 1 |
|
| | highlights | | |
| 100.00% | Emotion telling (show vs. tell) | Target: ≤3% sentences with emotion telling | | emotionTells | 0 | | narrationSentences | 216 | | matches | (empty) | |
| 100.00% | Filter word density | Target: ≤3% sentences with filter/hedge words | | filterCount | 2 | | hedgeCount | 0 | | narrationSentences | 216 | | filterMatches | | | hedgeMatches | (empty) | |
| 100.00% | Gibberish response detection | Target: ≤1% gibberish-like sentences (hard fail if a sentence exceeds 800 words) | | analyzedSentences | 216 | | gibberishSentences | 0 | | adjustedGibberishSentences | 0 | | longSentenceCount | 0 | | runOnParagraphCount | 0 | | giantParagraphCount | 0 | | wordSaladCount | 0 | | repetitionLoopCount | 0 | | controlTokenCount | 0 | | maxSentenceWordsSeen | 39 | | ratio | 0 | | matches | (empty) | |
| 100.00% | Markdown formatting overuse | Target: ≤5% words in markdown formatting | | markdownSpans | 0 | | markdownWords | 0 | | totalWords | 2331 | | ratio | 0 | | matches | (empty) | |
| 100.00% | Missing dialogue indicators (quotation marks) | Target: ≤10% speech attributions without quotation marks | | totalAttributions | 2 | | unquotedAttributions | 0 | | matches | (empty) | |
| 100.00% | Name drop frequency | Target: ≤1.0 per-name mentions per 100 words | | totalMentions | 67 | | wordCount | 2331 | | uniqueNames | 25 | | maxNameDensity | 0.64 | | worstName | "Silas" | | maxWindowNameDensity | 1.5 | | worstWindowName | "Silas" | | discoveredNames | | Carter | 1 | | Dad | 2 | | Cardiff | 1 | | London | 3 | | Brendan | 2 | | Yu-Fei | 1 | | Cheung | 1 | | Empress | 1 | | Malphora | 1 | | Aurora | 11 | | England | 1 | | Silas | 15 | | Blackwood | 2 | | Spymaster | 1 | | Prague | 2 | | Tuesday | 1 | | Soho | 3 | | Evan | 4 | | Berlin | 1 | | Si | 1 | | Europe | 1 | | Rory | 2 | | Raven | 1 | | Nest | 1 | | You | 7 |
| | persons | | 0 | "Carter" | | 1 | "Dad" | | 2 | "Brendan" | | 3 | "Yu-Fei" | | 4 | "Cheung" | | 5 | "Aurora" | | 6 | "Silas" | | 7 | "Blackwood" | | 8 | "Evan" | | 9 | "Rory" | | 10 | "Nest" | | 11 | "You" |
| | places | | 0 | "Cardiff" | | 1 | "London" | | 2 | "England" | | 3 | "Prague" | | 4 | "Soho" | | 5 | "Berlin" | | 6 | "Europe" | | 7 | "Raven" |
| | globalScore | 1 | | windowScore | 1 | |
| 100.00% | Narrator intent-glossing | Target: ≤2% narration sentences with intent-glossing patterns | | analyzedSentences | 173 | | 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 | 2331 | | matches | (empty) | |
| 100.00% | Overuse of "that" (subordinate clause padding) | Target: ≤2% sentences with "that" clauses | | thatCount | 0 | | totalSentences | 216 | | matches | (empty) | |
| 75.69% | Paragraph length variance | Target: CV ≥0.5 for paragraph word counts | | totalParagraphs | 56 | | mean | 41.63 | | std | 17.27 | | cv | 0.415 | | sampleLengths | | 0 | 48 | | 1 | 49 | | 2 | 4 | | 3 | 21 | | 4 | 24 | | 5 | 25 | | 6 | 50 | | 7 | 30 | | 8 | 48 | | 9 | 68 | | 10 | 28 | | 11 | 16 | | 12 | 41 | | 13 | 35 | | 14 | 46 | | 15 | 15 | | 16 | 61 | | 17 | 42 | | 18 | 47 | | 19 | 47 | | 20 | 30 | | 21 | 58 | | 22 | 23 | | 23 | 63 | | 24 | 41 | | 25 | 57 | | 26 | 35 | | 27 | 45 | | 28 | 19 | | 29 | 35 | | 30 | 13 | | 31 | 42 | | 32 | 17 | | 33 | 31 | | 34 | 32 | | 35 | 42 | | 36 | 42 | | 37 | 31 | | 38 | 33 | | 39 | 2 | | 40 | 61 | | 41 | 74 | | 42 | 62 | | 43 | 70 | | 44 | 54 | | 45 | 48 | | 46 | 35 | | 47 | 42 | | 48 | 49 | | 49 | 50 |
| |
| 100.00% | Passive voice overuse | Target: ≤2% passive sentences | | passiveCount | 1 | | totalSentences | 216 | | matches | | |
| 100.00% | Past progressive (was/were + -ing) overuse | Target: ≤2% past progressive verbs | | pastProgressiveCount | 0 | | totalVerbs | 386 | | matches | (empty) | |
| 100.00% | Em-dash & semicolon overuse | Target: ≤2% sentences with em-dashes/semicolons | | emDashCount | 0 | | semicolonCount | 0 | | flaggedSentences | 0 | | totalSentences | 216 | | ratio | 0 | | matches | (empty) | |
| 100.00% | Purple prose (modifier overload) | Target: <4% adverbs, <2% -ly adverbs, no adj stacking | | wordCount | 2337 | | adjectiveStacks | 0 | | stackExamples | (empty) | | adverbCount | 79 | | adverbRatio | 0.03380402225074882 | | lyAdverbCount | 17 | | lyAdverbRatio | 0.007274283269148481 | |
| 100.00% | Repeated phrase echo | Target: ≤20% sentences with echoes (window: 2) | | totalSentences | 216 | | echoCount | 0 | | echoWords | (empty) | |
| 100.00% | Sentence length variance | Target: CV ≥0.4 for sentence word counts | | totalSentences | 216 | | mean | 10.79 | | std | 6.49 | | cv | 0.601 | | sampleLengths | | 0 | 6 | | 1 | 11 | | 2 | 18 | | 3 | 13 | | 4 | 7 | | 5 | 8 | | 6 | 15 | | 7 | 19 | | 8 | 2 | | 9 | 2 | | 10 | 12 | | 11 | 9 | | 12 | 14 | | 13 | 10 | | 14 | 12 | | 15 | 6 | | 16 | 7 | | 17 | 9 | | 18 | 3 | | 19 | 6 | | 20 | 8 | | 21 | 3 | | 22 | 21 | | 23 | 10 | | 24 | 6 | | 25 | 14 | | 26 | 10 | | 27 | 4 | | 28 | 3 | | 29 | 3 | | 30 | 9 | | 31 | 3 | | 32 | 16 | | 33 | 3 | | 34 | 18 | | 35 | 11 | | 36 | 10 | | 37 | 26 | | 38 | 1 | | 39 | 9 | | 40 | 6 | | 41 | 3 | | 42 | 9 | | 43 | 7 | | 44 | 9 | | 45 | 4 | | 46 | 7 | | 47 | 16 | | 48 | 8 | | 49 | 2 |
| |
| 57.10% | Sentence opener variety | Target: ≥60% unique sentence openers | | consecutiveRepeats | 10 | | diversityRatio | 0.37037037037037035 | | totalSentences | 216 | | uniqueOpeners | 80 | |
| 81.70% | Adverb-first sentence starts | Target: ≥3% sentences starting with an adverb | | adverbCount | 5 | | totalSentences | 204 | | matches | | 0 | "Then the words turned to" | | 1 | "Always preferred that to Aurora." | | 2 | "Too formal, you said." | | 3 | "Mostly to myself at three" | | 4 | "Instead I talked about operations" |
| | ratio | 0.025 | |
| 96.47% | Pronoun-first sentence starts | Target: ≤30% sentences starting with a pronoun | | pronounCount | 63 | | totalSentences | 204 | | matches | | 0 | "She scanned the empty tables" | | 1 | "His hazel eyes narrowed against" | | 2 | "He gripped the edge of" | | 3 | "She crossed the floor in" | | 4 | "You look like the stories" | | 5 | "His limp showed as he" | | 6 | "He poured whiskey with steady" | | 7 | "She took a slow swallow" | | 8 | "I got out." | | 9 | "His gaze flicked to her" | | 10 | "Her black hair hung straight" | | 11 | "Your dad mentioned the name" | | 12 | "I should've written." | | 13 | "She turned the glass in" | | 14 | "I stopped answering calls from" | | 15 | "You know the place?" | | 16 | "He gave a short nod" | | 17 | "His limp became more obvious" | | 18 | "I eat there twice a" | | 19 | "It faded before it settled." |
| | ratio | 0.309 | |
| 67.84% | Subject-first sentence starts | Target: ≤72% sentences starting with a subject | | subjectCount | 160 | | totalSentences | 204 | | matches | | 0 | "Aurora shoved the heavy door" | | 1 | "The green neon sign buzzed" | | 2 | "Maps and black-and-white photographs lined" | | 3 | "She scanned the empty tables" | | 4 | "Silas straightened at the sight" | | 5 | "His hazel eyes narrowed against" | | 6 | "The neatly trimmed beard had" | | 7 | "He gripped the edge of" | | 8 | "She crossed the floor in" | | 9 | "The scar on her left" | | 10 | "You look like the stories" | | 11 | "His limp showed as he" | | 12 | "The knee gave a faint" | | 13 | "He poured whiskey with steady" | | 14 | "The field took more than" | | 15 | "Prague in ninety-eight." | | 16 | "Aurora wrapped her fingers around" | | 17 | "The liquid caught the low" | | 18 | "She took a slow swallow" | | 19 | "Law lasted about as long" |
| | ratio | 0.784 | |
| 100.00% | Subordinate conjunction sentence starts | Target: ≥2% sentences starting with a subordinating conjunction | | subConjCount | 5 | | totalSentences | 204 | | matches | | 0 | "Now I pour drinks instead" | | 1 | "Now I live upstairs from" | | 2 | "Now we both serve people" | | 3 | "Now I stand still and" | | 4 | "Even if we've both become" |
| | ratio | 0.025 | |
| 100.00% | Technical jargon density | Target: ≤6% sentences with technical-jargon patterns | | analyzedSentences | 106 | | technicalSentenceCount | 5 | | matches | | 0 | "The bright blue eyes carried a flat calm that hadn't been there in the girl who once begged for tales of secret rooms and coded messages." | | 1 | "I stopped answering calls from anyone who knew me as that girl with the law books." | | 2 | "The great Si Blackwood who knew every back door in Europe." | | 3 | "Now we both serve people who don't know our real names." | | 4 | "The limp carried him toward the back as the clock advanced, each tick marking another small surrender to the years that had changed his old friend's daughter in…" |
| |
| 100.00% | Useless dialogue additions | Target: ≤5% dialogue tags with trailing filler fragments | | totalTags | 0 | | uselessAdditionCount | 0 | | matches | (empty) | |
| 100.00% | Dialogue tag variety (said vs. fancy) | Target: ≤10% fancy dialogue tags | |