buvard 0.0.0-stage → 0.1.0

package/README.md

@@ -1,3 +1,96 @@
-# Temporary Holding Version
+# buvard
 
-This version is a temporary placeholder for this package. An operational version to replace this has been submitted for review and is awaiting a staged release.
+**L'anti-slop qui tache.** Blotting paper for AI-flavored prose: a weighted,
+multi-score linter for French and English slop patterns. It absorbs what
+baves — the empty openers, the inflated adjectives, the "Ce n'est pas X,
+c'est Y" pivots, the tirets cadratins by the dozen.
+
+One catalog, clustered from five anti-slop sources (two French skills, two
+bilingual frameworks, one EN framework), scored by a ratified barème, and
+tested. Zero config, by design — like black, the formatting is the tool's;
+the words stay the author's.
+
+![Anatomie d'une bavure — one slop sentence, three hits, 100 → 77](docs/anatomie-du-slop.svg)
+
+```bash
+git clone https://git.nech.pl/pln/buvard.git && cd buvard
+sudo ln -sf "$PWD/bin/buvard.js" /usr/local/bin/buvard
+
+buvard check posts/            # scores + grouped hits, per file
+buvard check article.md --json # machine-readable, for pipelines
+```
+
+## The barème
+
+Score starts at 100. A lexique hit costs 5, a structure hit 8, servility or
+meta-commentary 10. Punctuation runs on budgets: one tiret cadratin per 500
+words, one `!` per 1000 — each excess occurrence costs 5.
+
+- **100–90** — publish.
+- **80–95** — real hits only; fix what reads wrong.
+- **<80** — rewrite pass, then re-run. The score is a smoke alarm, not a grade.
+
+Every hit is a review candidate, not a verdict. Never touch: the author's
+ideas, data, deliberate jargon, quoted speech, or the voice.
+
+## Scope and honesty
+
+buvard is a **style linter**, not an AI detector. It finds statistically
+over-represented patterns; it cannot and does not claim to detect machine
+authorship. Its catalog is deliberately regex-shaped (high precision, low
+recall): règle de trois, synonyme acrobatique and sentence-length monotony
+stay manual checks. See [Science and references](#science-and-references)
+for the evidence behind the catalog.
+
+## Agents
+
+Agents run buvard three ways, in order of friction:
+
+```bash
+npx buvard check .            # after npm publish — no install at all
+node bin/buvard.js check .    # from a clone — zero dependencies, Node >= 18
+buvard check .                # global symlink (see install above)
+```
+
+`AGENTS.md` carries the project handoff for any agent; `SKILL.md` packages the
+catalog as an agent skill: load it, run the linter, apply the findings by hand.
+Hits are review candidates, never verdicts — the agent edits prose like an
+editor, not like a filter.
+
+## Science and references
+
+The catalog is data with provenance, and the barème is a ratified decision —
+both are written down:
+
+- **[docs/research.md](docs/research.md)** — the clustering study: five
+  anti-slop sources (fr1, fr2, étape, no-slop, ui-sk), what each contributes,
+  and how the shared patterns became the catalog.
+- **[docs/research-litreview.md](docs/research-litreview.md)** — the
+  literature review on AI-text detection, and the **ten falsifiable
+  recommendations** that will become catalog v3. Read it before touching
+  the catalog.
+
+Core citations, one line each — the lit review holds the full bibliography:
+
+- Liang et al., 2023 — [GPT detectors are biased against non-native English writers](https://arxiv.org/abs/2304.02819): a 61.22% false-positive rate on TOEFL essays. The case for staying a linter.
+- [MULTITuDE](https://arxiv.org/abs/2310.13606) — the large multilingual detection benchmark omits French; [INRIA's French ChatGPT-detection study](https://arxiv.org/abs/2306.05871) shows French is open territory.
+- [GLTR](https://arxiv.org/abs/1906.04043), [DetectGPT](https://arxiv.org/abs/2301.11305), [Binoculars](https://arxiv.org/abs/2401.12070) — the detection toolbox; each degrades on short or paraphrased text.
+- Sadasivan et al., 2024 — [paraphrase attacks collapse detection](https://arxiv.org/abs/2303.11156) on ~300-token passages. Why buvard never plays that game.
+- Kobak et al., 2024 — [excess vocabulary](https://arxiv.org/abs/2403.07183) as a population-level AI marker. A cousin of our lexique.
+- Sibling linters — [proselint](https://github.com/amperser/proselint) and [Vale](https://docs.vale.sh/) are the ancestors; [Wikipedia's "Signs of AI writing"](https://en.wikipedia.org/wiki/Wikipedia:Signs_of_AI_writing) is the shared folklore.
+
+**Methodology, going forward (catalog v3):** every new rule must pass a
+false-positive gate — under 2% per rule on a pre-2022 Gutenberg corpus,
+human-written by construction. Model-generated text is used only to measure
+sensitivity, never to admit a rule; a catalog tuned on one model's slop
+learns that model's fingerprints.
+
+## Skill
+
+`SKILL.md` ships the catalog as an agent skill: load it, run the linter,
+apply the findings by hand. The remedies are content, not code.
+
+## License
+
+GPL-3.0-only — see [LICENSE](LICENSE). The name: a buvard is blotting paper;
+it exists because l'encre bave.

package/SKILL.md

@@ -0,0 +1,34 @@
+---
+name: buvard
+description: Use when writing, editing or reviewing French or English prose that must not read as AI-generated — blog posts, docs, newsletters, READMEs — or before publishing any text an agent drafted. Runs the buvard linter and applies its catalog by hand.
+---
+
+# buvard — l'anti-slop qui tache
+
+Blotting paper for AI-flavored prose. A weighted catalog of slop patterns
+clustered from five anti-slop sources, plus punctuation budgets
+(1 tiret cadratin per 500 words, 1 `!` per 1000).
+
+## Run it
+
+```bash
+buvard check <files|dirs>     # scores + grouped hits
+buvard check . --json         # machine-readable
+```
+
+## Apply the findings
+
+1. Every hit is a **review candidate**, not a verdict. Read the sentence.
+2. The remedy is always **content, not code**: cut the opener, name the
+   actor, say the second half directly, replace the adjective with the fact
+   that justifies it. Repetition of the right word beats an elegant synonym.
+3. Never touch: the author's ideas, data, deliberate technical jargon,
+   quoted speech, or the voice. Two `!` in a joyful 137-word post is not slop.
+4. Known regex gaps, check by eye: règle de trois, synonyme acrobatique,
+   sentence-length monotony, parataxis chains.
+
+## Score interpretation
+
+- 100–90: publish.
+- 80–95: real hits only; fix what reads wrong.
+- <80: rewrite pass, then re-run. Score is a smoke alarm, not a grade.

package/bin/buvard.js

@@ -0,0 +1,68 @@
+#!/usr/bin/env node
+/**
+ * buvard — l'anti-slop qui tache. Blotting paper for AI-flavored prose.
+ *
+ * buvard check <files|dirs> [--json]   run the catalog over markdown/text
+ * buvard score <file>                  single-file score, machine-readable
+ *
+ * Scoring starts at 100; each hit costs its weight (lexique 5, structure 8,
+ * servile 10; punctuation charges per occurrence past the budget: one tiret
+ * cadratin per 500 words, one exclamation per 1000). Every hit is a review
+ * candidate, not a verdict — the author's voice always wins.
+ *
+ * Known gaps, deliberately manual: règle de trois, synonyme acrobatique and
+ * sentence-length variance need document-level analysis that regex fakes
+ * badly. Read the text too.
+ */
+
+import fs from 'fs';
+import path from 'path';
+import { CATALOG_VERSION, lint, bodyOf } from '../lib/index.mjs';
+
+const TEXT_EXT = /\.(md|txt|mdx)$/i;
+
+function collect(target, files) {
+  if (TEXT_EXT.test(target) && fs.existsSync(target)) return files.push(path.resolve(target));
+  if (fs.existsSync(target) && fs.statSync(target).isDirectory()) {
+    for (const f of fs.readdirSync(target)) {
+      const p = path.join(target, f);
+      if (fs.statSync(p).isDirectory()) collect(p, files);
+      else if (TEXT_EXT.test(f)) files.push(path.resolve(p));
+    }
+    return;
+  }
+  console.error(`buvard: no file or directory for "${target}"`);
+}
+
+function main(argv) {
+  const json = argv.includes('--json');
+  const targets = argv.filter((a) => !a.startsWith('--'));
+  const files = [];
+  if (targets.length > 0) targets.forEach((t) => collect(t, files));
+  else collect('.', files);
+
+  const results = files.map((file) => ({ file, ...lint(bodyOf(fs.readFileSync(file, 'utf8'))) }));
+
+  if (json) {
+    console.log(JSON.stringify({ buvard: CATALOG_VERSION, results }, null, 1));
+    return 0;
+  }
+  for (const r of results) {
+    const grouped = {};
+    for (const h of r.hits) (grouped[h.id] ??= []).push(h);
+    console.log(`\n== ${path.relative(process.cwd(), r.file)} — ${r.score}/100 (${r.words} mots)`);
+    for (const [id, list] of Object.entries(grouped)) {
+      console.log(`  [${id}] ×${list.length} (-${list[0].w} chacun)`);
+      for (const h of list.slice(0, 4)) console.log(`      ${h.at}`);
+    }
+  }
+  console.log(`\n---- buvard v${CATALOG_VERSION} ----`);
+  for (const r of results) console.log(`${String(r.score).padStart(3)}/100  ${path.relative(process.cwd(), r.file)}`);
+  const worst = results.filter((r) => r.hits.length > 0).sort((a, b) => a.score - b.score)[0];
+  if (worst) console.log(`priorité: ${path.relative(process.cwd(), worst.file)} (${worst.score}/100)`);
+  return 0;
+}
+
+if (process.argv[1] && path.resolve(process.argv[1]) === path.resolve(import.meta.url.replace('file://', ''))) {
+  process.exit(main(process.argv.slice(2)));
+}

package/docs/anatomie-du-slop.svg

@@ -0,0 +1,76 @@
+<svg xmlns="http://www.w3.org/2000/svg" viewBox="0 0 880 470" role="img" aria-labelledby="t d">
+  <title id="t">Anatomie d'une bavure</title>
+  <desc id="d">One slop sentence dissected: buvard's catalog flags a binary contrast (−8), a servile promise (−10) and a second tiret cadratin beyond budget (−5); the score falls from 100 to 77.</desc>
+  <defs>
+    <style>
+      .ink  { fill:#1f2a33; }
+      .rouge{ fill:#a63d2f; }
+      .muted{ fill:#7a6f5d; }
+      .serif{ font-family: Georgia, 'Times New Roman', serif; }
+      .sans { font-family: 'Avenir Next', 'Helvetica Neue', Arial, sans-serif; }
+      .line { stroke:#1f2a33; fill:none; stroke-linecap:round; }
+      .rouged { stroke:#a63d2f; fill:none; stroke-linecap:round; }
+    </style>
+  </defs>
+
+  <!-- paper -->
+  <rect width="880" height="470" fill="#faf5ea"/>
+
+  <!-- title -->
+  <text x="90" y="62" class="sans ink" font-size="15" letter-spacing="6">ANATOMIE D'UNE BAVURE</text>
+  <path d="M90 74 q 60 4 120 1 t 122 -1" class="line" stroke-width="1.4"/>
+  <text x="790" y="62" class="sans muted" font-size="11" text-anchor="end" letter-spacing="2">buvard</text>
+
+  <!-- the sentence, segment by segment so the boxes align exactly -->
+  <g class="serif ink" font-size="21">
+    <text x="90"  y="140" textLength="24"  lengthAdjust="spacingAndGlyphs">«&#160;</text>
+    <text x="114" y="140" textLength="222" lengthAdjust="spacingAndGlyphs">Ce n'est pas un outil,</text>
+    <text x="336" y="140" textLength="205" lengthAdjust="spacingAndGlyphs">c'est un compagnon</text>
+    <text x="90"  y="182" textLength="100" lengthAdjust="spacingAndGlyphs">de voyage</text>
+    <text x="196" y="182" textLength="26"  lengthAdjust="spacingAndGlyphs">&#160;—&#160;</text>
+    <text x="228" y="182" textLength="92"  lengthAdjust="spacingAndGlyphs">et il va</text>
+    <text x="326" y="182" textLength="200" lengthAdjust="spacingAndGlyphs">vous changer la vie</text>
+    <text x="532" y="182" textLength="52"  lengthAdjust="spacingAndGlyphs">&#160;!&#160;»</text>
+  </g>
+
+  <!-- red-pencil boxes around the three hits -->
+  <rect x="108" y="118" width="234" height="33" rx="7" class="rouged" stroke-width="1.6" stroke-dasharray="1 0"/>
+  <rect x="191" y="160" width="36"  height="33" rx="7" class="rouged" stroke-width="1.6"/>
+  <rect x="320" y="160" width="212" height="33" rx="7" class="rouged" stroke-width="1.6"/>
+
+  <!-- leader lines -->
+  <path d="M225 151 Q 200 205 150 236" class="line" stroke-width="1" stroke-dasharray="3 4"/>
+  <circle cx="150" cy="236" r="2.2" class="rouge" stroke="none"/>
+  <path d="M209 193 Q 250 240 310 284" class="line" stroke-width="1" stroke-dasharray="3 4"/>
+  <circle cx="310" cy="284" r="2.2" class="rouge" stroke="none"/>
+  <path d="M426 193 Q 480 240 533 296" class="line" stroke-width="1" stroke-dasharray="3 4"/>
+  <circle cx="533" cy="296" r="2.2" class="rouge" stroke="none"/>
+
+  <!-- the three verdicts -->
+  <g class="serif">
+    <text x="90" y="266" class="rouge" font-size="17" font-weight="bold">contraste binaire <tspan class="rouge">&#8722;8</tspan></text>
+    <text x="90" y="286" class="muted" font-size="12.5" font-style="italic">« pas X, c'est Y » — la bascule creuse</text>
+
+    <text x="320" y="300" class="rouge" font-size="17" font-weight="bold">tiret cadratin <tspan class="rouge">&#8722;5</tspan></text>
+    <text x="320" y="320" class="muted" font-size="12.5" font-style="italic">budget : un seul pour 500 mots</text>
+
+    <text x="545" y="312" class="rouge" font-size="17" font-weight="bold">servile <tspan class="rouge">&#8722;10</tspan></text>
+    <text x="545" y="332" class="muted" font-size="12.5" font-style="italic">la promesse d'émerveillement</text>
+  </g>
+
+  <!-- the gauge: 0–100, threshold at 80, ink fill to 77 -->
+  <g>
+    <text x="90" y="368" class="sans ink" font-size="12" letter-spacing="3">SCORE</text>
+    <line x1="90" y1="400" x2="790" y2="400" class="line" stroke-width="1.2"/>
+    <line x1="90" y1="400" x2="646" y2="400" class="line" stroke-width="7"/>
+    <!-- threshold tick at 80 : 90 + 0.80*700 = 650 -->
+    <line x1="650" y1="390" x2="650" y2="410" class="rouged" stroke-width="1.6"/>
+    <text x="650" y="426" class="rouge sans" font-size="10.5" text-anchor="middle" letter-spacing="1">SEUIL 80</text>
+    <text x="790" y="426" class="muted sans" font-size="10.5" text-anchor="end">100</text>
+    <text x="90" y="426" class="muted sans" font-size="10.5">0</text>
+    <text x="646" y="384" class="rouge serif" font-size="26" font-weight="bold" text-anchor="end">77</text>
+    <text x="676" y="368" class="muted serif" font-size="13" font-style="italic">réécrire, puis re-passer</text>
+  </g>
+
+  <text x="440" y="458" class="sans muted" font-size="10" text-anchor="middle" letter-spacing="2">HITS ARE REVIEW CANDIDATES, NEVER VERDICTS — THE AUTHOR'S VOICE ALWAYS WINS</text>
+</svg>

package/docs/research-litreview.md

@@ -0,0 +1,282 @@
+# buvard: literature review
+
+Literature review supporting buvard, a regex + heuristic prose linter for AI-slop
+patterns in French and English. Written 2026-11-07. Every non-obvious claim carries a
+URL; anything we could not verify is marked [inferred] or [unverified]. This document is
+itself written under buvard's own rules: no em-dash clause separators, no contrast
+constructions, no rule-of-three defaults.
+
+## 1. Statistical markers: what stylometry actually measures
+
+Formal detection research works at the token level, with probability scores, while
+buvard works at the surface level, with regex and counts. The two literatures agree on
+one thing that matters for buvard: short texts defeat everything.
+
+The canonical systems and their numbers:
+
+| System / metric | What it measures | Reported result | Source |
+|---|---|---|---|
+| GLTR (token rank) | position of each token in the model's ranked vocabulary; human text draws from ranks 10-1000 far more than AI text | raised human judges' fake-text accuracy from 54% to 72% on GPT-2 output, with no training | https://arxiv.org/abs/1906.04043 |
+| DetectGPT (curvature) | negative curvature of the log-probability surface under random perturbations | AUROC on fake-news detection rose from 0.81 to 0.95; requires logprobs from the source model itself | https://arxiv.org/abs/2301.11305 |
+| Binoculars (cross-perplexity) | ratio of perplexities of two closely related models on the same text | over 90% recall on ChatGPT samples at 0.01% false positive rate | https://arxiv.org/abs/2401.12070 |
+| Recursive paraphrase attack | paraphrase round-trips push text back toward the human perplexity band | detection rates collapse on ~300-token passages while text quality barely drops | https://arxiv.org/abs/2303.11156 |
+
+Three limits recur across these papers and they define buvard's lane:
+
+1. **Token-level scores need the model's logits.** DetectGPT needs the source model's
+   own logprobs, Binoculars needs two local models, GLTR needs the vocabulary ranking.
+   A regex linter cannot have any of these. That is a feature here: the catalog catches
+   surface evidence, and it says so.
+2. **Short texts are unreliable for everyone.** Sadasivan et al. test on ~300-token
+   passages and show paraphrasing alone destroys detection. Any per-paragraph verdict
+   from a stylometric score on fewer than ~150 words is noise. buvard's weighted scores
+   should aggregate per document, per 500-word window, and report counts, never a
+   probability of AI authorship.
+3. **Detectors discriminate against human writers as much as machines.** Seven commercial GPT
+   detectors misclassified on average 61.22% of 91 TOEFL essays by non-native writers
+   as AI-generated; 97.80% were flagged by at least one detector. Prompting to sound
+   "more like a native speaker" cut the false positive rate from 61.22% to 11.77%,
+   which pins the cause on word choice: common-word-heavy, low-perplexity prose reads
+   as machine. Source: https://arxiv.org/abs/2304.02819 (published in Patterns:
+   https://www.cell.com/patterns/fulltext/S2666-3899(23)00130-7). This is the single
+   strongest argument for buvard to stay a style linter and never an authorship judge.
+
+On the specific surface metrics writers talk about:
+
+- **Perplexity and burstiness.** GPTZero popularized "perplexity" (predictability) and
+  "burstiness" (sentence-to-sentence variance) as its two scores; the company's own
+  learning pages define burstiness as variation in sentence structure and length
+  [inferred URL: https://gptzero.me/learn/, we could not fetch it cleanly]. The
+  direction is consistent everywhere: AI text has flatter sentence-length variance.
+  Note the fairness trap: lower variance is also a hallmark of trained technical
+  writing, so buvard should flag flat rhythm as a structural note, at low weight.
+- **Function-word rates and lexical diversity.** Liang et al.'s corpus method
+  (frequency of common tokens against expert-written reference texts) found a
+  "decrease in variation of linguistic features" in LLM-modified text and estimated
+  6.5-16.9% of AI-conference peer-review text as substantially LLM-modified.
+  Source: https://arxiv.org/abs/2403.07183. Kobak et al.'s excess-vocabulary analysis
+  of 15M PubMed abstracts (2010-2024) showed an abrupt step increase in specific style
+  words the month ChatGPT shipped, and a floor of at least 13.5% of 2024 abstracts
+  LLM-processed, up to 40% in some subcorpora. Source:
+  https://arxiv.org/abs/2406.07016 (published version in Science Advances, 2025).
+  For buvard: relative word frequency shifts are the most robust, hardest-to-avoid
+  signal, and they are exactly what a frequency-banded lexicon encodes.
+- **Repetition metrics.** Low type-token ratio and repeated sentence openers are
+  standard stylometric evidence; the excess-vocabulary paper above is the strongest
+  published version of the repetition claim (same words, everywhere, all at once).
+  Sentence-opener repetition is also the top complaint in every practitioner catalog
+  reviewed in section 4.
+
+
+## 2. The contrast-construction family and its cousins
+
+**"Not X, but Y" / "Ce n'est pas X, c'est Y".** Wikipedia's "Signs of AI writing"
+maintainer page lists the construction as a top indicator, in the forms "not just X,
+but also Y" and "serves as / stands as / marks / functions as / operates as /
+represents [a]": https://en.wikipedia.org/wiki/Wikipedia:Signs_of_AI_writing. The
+page's own caveat matters and buvard should quote it in its docs: these are "potential
+signs", since human authors use the same styles constantly, especially in blogs and
+list-style articles. hardikpandya's stop-slop skill independently catalogs "binary
+contrasts" as an AI tell: https://github.com/hardikpandya/stop-slop. On the French
+side, both anti-slop-fr skills flag structural contrast and false-agentivity frames
+("les chiffres parlent d'eux-mêmes", "plongeons dans"): see section 3.
+
+The craft consensus, reconstructed from these catalogs: the antithesis is classical
+rhetoric and legitimate in moderation; the tell is frequency and sameness. LLMs reach
+for the construction as a default compression move, so one instance per document is
+style and five per 500 words is slop. For buvard this argues a density threshold with
+a low per-instance penalty, not a per-hit kill.
+
+**The rising "X, so Y" pattern.** This is the "It's a great tool, so..." hedge-plus-
+pivot: a bland claim followed by "so" + grandiose consequence ("so you can focus on
+what matters"). We found no published linguistic study of it [unverified]; it appears
+in practitioner catalogs, including the 20-pattern creatoreconomy skill buvard already
+ingests. Treat the sourcing as folklore for now: flag at low weight until a second
+independent source appears.
+
+**Em dash.** Wikipedia's signs page documents "Overuse of em dashes" with an explicit
+date range, 2022 to September 2026, tying the spike to LLM output
+(https://en.wikipedia.org/wiki/Wikipedia:Signs_of_AI_writing). The 2025 press wave on
+this is real but unreachable from this environment, so specific articles (Merriam-
+Webster wordplay, Guardian comment pieces, New York Magazine's em-dash coverage) and
+the reported Microsoft Word autocorrect incident that auto-inserted em dashes and was
+reverted after user backlash are marked [unverified] here. buvard's own budget
+approach (a cap per 500 words instead of a ban) matches the actual debate: writers
+with strong opinions on em dashes are the loudest false-positive victims, and
+hand-written English prose legitimately uses them. French differs: French typography
+uses the tiret for dialogue and parenthetical incises far less in expository prose, so
+the FR budget should be tighter.
+
+**Rule of three.** The tricolon is a classical figure, and Wikipedia's signs page
+lists "Rule of three" as an AI indicator without claiming any single tricolon is bad.
+The practitioner consensus again is density: a consistent pattern of A, B, and C
+triplets in consecutive paragraphs is the tell. buvard should detect the repetition
+across a window (three or more tricolons in 500 words) rather than flag each one, and
+exempt dialogue, lists, and comedic writing where the triple is the joke.
+
+**Design rule extracted from this whole section.** For every rhetorical device in the
+family (contrast, tricolon, dash, so-pivot), the literature and the catalogs converge
+on the same shape: legitimate once, slop when uniform. Encode per-instance penalties
+near zero and window-density penalties that bite.
+
+
+## 3. French-specific AI tells: thinner coverage, different typography
+
+**The practitioner layer is where FR is strongest.** The two anti-slop-fr skills are
+the main public French catalogs:
+
+- JulienSnsnCode/anti-slop-fr (https://github.com/JulienSnsnCode/anti-slop-fr):
+  organizes the problem into 13 lexicon categories and 13 structural patterns, flags
+  over-represented words ("mots surreprésentés"), ambiguous turns, predictable
+  structures such as "dans un monde en constante évolution", English calques
+  (« plongeons dans »), and false agentivity (« les chiffres parlent d'eux-mêmes »).
+  It grades text 1 to 10 on Franchise, Rythme, Confiance, Naturel, Densité and
+  rewrites under 35/50, which is an LLM-judge pipeline, unlike buvard's deterministic
+  scoring.
+- captain-marketing/anti-slop-fr (https://github.com/captain-marketing/anti-slop-fr):
+  two main families, "Formules creuses" (empty hyperboles such as « révolutionnaire »)
+  and "Patterns structurels" (monotone rhythm, vague declaratives). It ships
+  identity/ voice profiles and examples/, i.e. it is a rewriting skill, again not a
+  linter.
+
+The letape-dapres.fr writing-unslop source could not be fetched from this environment;
+its contribution to buvard's catalog remains [unverified] pending a direct read.
+
+**The research layer barely covers French.** MULTITuDE, the largest multilingual
+machine-text benchmark, spans 74,081 texts in 11 languages and French is absent (ar,
+ca, cs, de, en, es, nl, pt, ru, uk, zh): https://arxiv.org/abs/2310.13606. The main
+French ChatGPT-detection study (INRIA, https://arxiv.org/abs/2306.05871, artifacts at
+https://gitlab.inria.fr/wantoun/robust-chatgpt-detection) built its French data by
+translating an English dataset, and its own conclusion is caution: in-domain accuracy
+held, out-of-domain and adversarial text exposed the detectors. So the only formal FR
+work is (a) translated, and (b) out-of-domain fragile. No published French excess-
+vocabulary analysis equivalent to Kobak et al. exists as far as we could find
+[unverified]. Conclusion: buvard's French catalog is not porting a mature literature,
+it is the literature, which is the project's real opportunity.
+
+**French slop vocabulary worth weighting (from the FR skills, cross-checked against
+EN patterns by analogy [inferred]):** « plonger dans », « à l'ère de », « en constante
+évolution », « il convient de noter », « il est important de noter », « crucial »,
+« essentiel », « révolutionnaire », « palpitant », « intrinsèquement », « en
+définitive », « à la fin de la journée », « les chiffres parlent d'eux-mêmes », «
+permet de », « grâce à » used as a hedge, and the English calques « débloquer » (au
+sens figuré), « donner un aperçu de ». These deserve frequency-band treatment like the
+EN lexicon rather than flat bans.
+
+**Typography is where FR detection differs structurally.** French orthography requires
+a non-breaking (or narrow no-break) space before double punctuation (; : ! ?) and
+inside guillemets « » : https://fr.wikipedia.org/wiki/Espace_ins%C3%A9cable (see also
+the Lexique des règles typographiques de l'Imprimerie nationale). LLM output is
+notoriously inconsistent here: thin or plain spaces slip in, and many LLM pipelines
+emit straight quotes instead of « ». buvard can add mechanical FR-only rules with
+almost zero false-positive risk: guillemet spacing, insécable before ; : ! ?, and
+straight-quote detection. These are the FR equivalent of the em dash, except they are
+genuinely wrong under French typographic convention, so they can carry a hard penalty
+instead of a stylistic note.
+
+**Coverage gap, quantified as far as the evidence allows:** English has the GLTR /
+DetectGPT / Binoculars literature, a commercial detector market, and dozens of
+practitioner catalogs. French has one translated-data detector study, one benchmark
+that skips it, two practitioner skills, and typography rules. buvard should say this
+on its README rather than pretend parity.
+
+
+## 4. The detector and linter landscape
+
+| Tool | Kind | What it actually does | Accuracy claims and FP profile |
+|---|---|---|---|
+| stop-slop (hardikpandya) | rules skill, MIT | an LLM agent skill, not a program: flags throat-clearing openers, business jargon, binary contrasts, passive voice; no benchmarks | none published; https://github.com/hardikpandya/stop-slop |
+| humanizer (blader) | LLM skill, MIT | a rewrite skill over "26 patterns" (staging instead of stating, rhythm by rule, inflation and borrowed authority, formatting by rule); a blind test had judges prefer its rewrite 16/16 | no detection accuracy claim; it is a rewriter; https://github.com/blader/humanizer |
+| deslop / unslop (Cursor ecosystem) | editor rules | could not verify from this environment: plausible slugs exist (cursor.directory returned 429, candidate GitHub repos 404) [unverified]; the pattern family is LLM-agent rules that strip slop from code comments and chat remnants | none |
+| proselint | rules linter, BSD | the ancestor of this genre: a spell-checker for style, rules for cliches and errors, explicitly "aggregate knowledge about best practices in writing" | no accuracy claim; it reports rules fired, not authorship; https://github.com/amperser/proselint |
+| Vale | rules linter, MIT | "brings code-like linting to prose": YAML-defined styles with regex and extension points, no model, explicitly not a detector | https://docs.vale.sh/ |
+| GPTZero | ML service | perplexity and burstiness scores; claims "99.9% Accuracy" while its own text warns no detector is 100% accurate and results "should not be used to punish or as the final verdict" | https://gptzero.me/ ; the 61.22% TOEFL false-positive result (section 1) includes detectors of this class |
+| Pangram | ML service | a "traditional language model architecture" classifier for text, image and video; claims "over 99% accurate" with a "1 in 10,000 False positive rate", "verified by researchers" at U Chicago and U Maryland | https://www.pangram.com/ ; the claim is vendor-published [inferred: independent replication details not accessible here] |
+| Originality.ai | ML service | "Most Accurate AI Detector in third-party studies"; no specific FP rate on its own front page, no visible appeal process | https://originality.ai/ |
+
+Reading the table as a competitive map for buvard:
+
+1. **The rules side of the market is agent skills, not linters.** stop-slop and
+   humanizer are prompts for LLMs to rewrite text. Nobody ships a deterministic,
+   score-weighted, CI-runnable prose linter focused on AI slop. proselint and Vale
+   are deterministic but generic-style, with no slop catalog and no French. This is
+   buvard's empty seat: Vale's architecture (YAML rules, regex, severities, exit
+   codes) pointed at a maintained FR+EN slop catalog.
+2. **ML detectors claim 99+ percent and disclaim everything.** GPTZero's own warning
+   against using results "to punish" is the confession the whole category shares, and
+   the peer-reviewed floor is the 61.22% false-positive rate on non-native writing
+   (section 1). Any tool that emits an authorship probability inherits this.
+3. **None of the rules tools has French.** proselint is English-only; Vale's official
+   styles are English-first [inferred]; both FR skills are prompts. buvard's FR rules
+   (section 3 typography rules in particular) have no direct competitor at all.
+4. **False-positive profile to design against.** The documented victims of AI
+   detection are non-native writers (arxiv 2304.02819) and, anecdotally across the
+   catalogs, autistic and ADHD writers whose prose is flatly organized or
+   formality-heavy [unverified, no published study found]. For a rules linter the
+   equivalent risk is harsh severity on ordinary rhetoric; buvard should default all
+   contrast-family rules to informational severity.
+
+
+## 5. Implications for buvard
+
+Ten recommendations, each stated so it can be falsified by a measurement:
+
+1. **Never claim AI detection.** Position buvard as a style linter that reports
+   surface slop evidence. No authorship probability, no "percent human". Test: the
+   output schema contains counts and weighted scores only. Grounds: 61.22% FPR on
+   non-native writers (section 1) and GPTZero's own anti-punishment disclaimer
+   (section 4).
+2. **Score the contrast family by density, not by instance.** Antithesis is
+   legitimate once and slop when uniform (section 2). Test: a pre-2022 Gutenberg
+   prose corpus should show the per-instance rule firing on classic essays at rates
+   indistinguishable from modern human prose, while the window rule separates the
+   corpora.
+3. **Promote FR typography rules to hard penalties.** Insécable before ; : ! ?,
+   guillemet spacing, straight quotes: mechanically wrong per French typographic law
+   (fr.wikipedia.org/wiki/Espace_ins%C3%A9cable), near-zero false-positive risk, and
+   no competitor has them. Test: FP rate on the Gutenberg FR corpus should be ~0
+   after normalizing Unicode.
+4. **Build the FR lexicon as frequency bands, not a flat list.** Tier 1 (always
+   slop: « il convient de noter », « à l'ère de », « plonger dans ») gets the full
+   weight; tier 2 (context-dependent: « crucial », « permet de ») fires only on
+   density. Grounds: the FR skills' two-tier structure (section 3) and the excess-
+   vocabulary finding that frequency shifts, not single words, are the robust signal
+   (arxiv 2406.07016).
+5. **Cap total lexicon penalty per document.** Common-word-heavy prose is exactly
+   what detectors punish unfairly (arxiv 2304.02819), so buvard must not reproduce
+   the failure in lint form. Test: a text with 40 tier-2 hits and nothing else
+   stays below the "fail" threshold.
+6. **Flag flat rhythm as an informational note.** Compute the coefficient of
+   variation of sentence lengths per 500-word window; flag low CV only in prose
+   genres where variance is expected (essays, narrative) and never in technical docs.
+   Thresholds are an open empirical question [inferred]; publish whatever the human
+   corpus shows.
+7. **Keep the "X, so Y" pivot at low weight and dated folklore status.** No
+   independent source found (section 2). Test for promotion: a second practitioner
+   catalog or an n-gram frequency study showing post-2022 growth. Until then it
+   cannot be promoted above -2.
+8. **Em-dash budgets per language.** EN 1 per 500 words (as shipped), FR tighter at
+   1 per 1000, both exempting dialogue lines. Grounds: the documented 2022-2026
+   em-dash spike is EN-centric (Wikipedia signs page), and FR prose uses tirets
+   mainly for dialogue and incises.
+9. **Rule-of-three fires on the window, never the first tricolon.** Test: flagging
+   only at three or more tricolons per 500 words should leave the Gutenberg corpus
+   nearly untouched.
+10. **Version the catalog like an API, with provenance on every rule.** Rules get
+    stable ids (`buvard.fr.lexique.007`), each carries `source` (URL + retrieval
+    date) and `severity`, the catalog semvers (adding a rule bumps minor, reweighting
+    bumps major since scores shift), and a machine-readable changelog. Grounds:
+    catalogs drift as LLMs change their tells; the em-dash range on the Wikipedia
+    page shows tells have lifetimes. Without provenance, a 2027 user cannot tell a
+    measured pattern from a 2024 hunch. This review's [unverified] marks are exactly
+    the metadata such a field would have made mandatory.
+
+**Summary of the position.** The statistical literature says detection itself is
+unreliable, biased against non-native writers, and defeated by paraphrasing; the
+commercial market sells 99% claims with disclaimers attached; the rules layer of the
+market is prompt skills with no determinism, no scores, and no French. A deterministic,
+score-weighted, provenance-tracked FR+EN style linter with a public false-positive
+test corpus occupies ground that nobody currently holds, and its honest claim ("your
+text reads like the slop everyone complains about") is the only claim in this space
+that survives contact with the evidence.
+

package/docs/research.md

@@ -0,0 +1,68 @@
+# Recherche anti-slop — cinq sources, une suite multiscore
+
+*2026-11-07. Contexte : l'axe slop de la suite prose (`yarn slop`, `scripts/slop.mjs`)
+est dérivé de ce document. L'axe grammaire reste `yarn prose` (LanguageTool).*
+
+## Les sources
+
+| # | Source | Nature | Ce qu'elle apporte d'unique |
+|---|---|---|---|
+| fr1 | [JulienSnsnCode/anti-slop-fr](https://github.com/JulienSnsnCode/anti-slop-fr) | skill FR, lexique 13 catégories + structures | le lexique FR le plus fin (kitsch narratif, faux poids, anti-copule) ; notation 5 dimensions /50 |
+| fr2 | [captain-marketing/anti-slop-fr](https://github.com/captain-marketing/anti-slop-fr) | skill FR, blacklist + `identity/VOICE.md` | le concept de **voix d'auteur** : les interdits génériques ne suffisent pas, il faut un profil de voix par auteur |
+| étape | [letape-dapres.fr writing-unslop](https://letape-dapres.fr/ressources/skills/writing-unslop) | skill bilingue FR+EN | listes « hard replacements » FR et EN ; scoring chiffré base 100 (-5/-8/-10) |
+| no-slop | [creatoreconomy.so, skill 20 patterns](https://creatoreconomy.so/p/use-my-no-ai-slop-skill-to-remove-20-ai-slop-patterns) | skill EN | la taxonomie la plus courte ; contrastes binaires, reveals au colon, fins faux-profondes |
+| ui-sk | [ui-skills.com anti-ai-slop-writing](https://www.ui-skills.com/skills/jalaalrd/anti-ai-slop-writing) | framework EN | les seuls **budgets de ponctuation quantifiés** (1 tiret cadratin / 500 mots, 1 ! / 1000) ; parataxe, hedging seesaw |
+
+Écosystème (flaviocopes [anti-slop](https://flaviocopes.com/anti-slop/)) : les skills
+concurrents s'appellent `unslop`, `stop-slop` (hardikpandya, MIT — l'ancêtre cité par fr1),
+`humanizer`, `deslop`. Le nôtre n'en remplace aucun : il les **compresse en linter exécutable**
+pour ce repo, là où ils sont des prompts de réécriture.
+
+## Le clustering — ce que les cinq disent ensemble
+
+Vote = nombre de sources qui portent la règle.
+
+| Cluster | Vote | Forme canonique |
+|---|---|---|
+| Tiret cadratin | **5/5** | jamais comme séparateur de proposition ; la phrase se coupe ou se colle |
+| Adjectifs gonflés | 5/5 | révolutionnaire, incontournable, robuste, seamless… le fait ou rien |
+| Règle de trois | 5/5 | deux éléments, ou un seul bien choisi |
+| Contraste binaire (« pas X, c'est Y ») | 5/5 | dire Y directement |
+| Adverbes de remplissage | 4/5 | vraiment, littéralement, fondamentalement → supprimer |
+| Ouvertures à vide | 4/5 | « Soyons honnêtes », « Dans un monde où » → commencer par le fond |
+| Anti-copule (constitue, serves as) | 4/5 | « est » presque toujours meilleur |
+| Attribution vague | 4/5 | nommer la source ou assumer |
+| Conclusions-bilans | 4/5 | finir sur un fait, pas un résumé |
+| Méta-commentaire / annonce de plan | 4/5 | le texte ne se raconte pas |
+| Fragments dramatiques | 3/5 | « Un mot. Un seul. » → phrase complète |
+| Question-réponse (« Le problème ? ») | 3/5 | phrase affirmative |
+| Faux poids (tournant, game-changer) | 3/5 | le chiffre ou l'exemple |
+| Servilité / artefacts collaboratifs | 3/5 | « N'hésitez pas à », « I hope this helps » → couper |
+
+## Ce que chaque source apporte en propre
+
+- **fr1** : kitsch narratif (« un silence vibrant ») ; synonyme acrobatique (répéter le
+  mot juste) ; la vérification « un adverbe en -ment ? ».
+- **fr2** : `VOICE.md` — la liste noire générique ne fait pas la voix ; les interdits
+  personnels et les expressions signature vivent à part. *À porter plus tard : un
+  `voice.md` par section du site.*
+- **étape** : les listes FR/EN prêtes pour regex ; le barème chiffré.
+- **no-slop** : reveals au colon (« The best part: ») ; fins faux-profondes.
+- **ui-sk** : budgets quantifiés (adoptés tels quels dans le linter) ; parataxe ;
+  « hedging seesaw » (prendre position).
+- **flaviocopes** : le slop de *code* (try/catch sur chemins sûrs, abstractions à un
+  appelant, casts `any`) — hors périmètre prose, à garder pour une revue code.
+
+## Le barème retenu (`scripts/slop.mjs`)
+
+Base 100. Lexique -5, structure -8, servilité/méta -10 ; ponctuation : -5 par
+occurrence au-delà du budget (1 tiret / 500 mots, 1 `!` / 1000 mots). Le catalogue
+est volontairement **haute précision** : chaque hit est un candidat relecture, pas un
+verdict. Le code compte les mots hors frontmatter.
+
+**Gaps assumés, à relire à l'œil** : règle de trois, synonyme acrobatique, variance de
+longueur de phrases, parataxe — l'analyse est document-niveaux, le regex la simule mal.
+
+**Ce que le linter ne touche jamais** (unanime des sources) : les idées, les données,
+le jargon technique assumé, et la voix de l'auteur. Un « ! » dans un post joyeux de
+137 mots n'est pas du slop.