@mrclrchtr/supi-flow 0.9.0 → 0.10.0

package/skills/supi-flow-slop-detect/scripts/slop-scan-vocab.ts

@@ -1,161 +0,0 @@
-#!/usr/bin/env -S pnpm exec jiti
-/**
- * Vocabulary marker scanner — scans markdown files for AI-prose vocabulary markers.
- *
- * Reads vocabulary definitions from `../references/vocabulary.json` — the single
- * source of truth shared with SKILL.md documentation tables.
- *
- * Usage:
- *   pnpm exec jiti scripts/slop-scan-vocab.ts <file> [<file>...]
- *
- * Cross-platform Node.js/TypeScript — runs wherever pi runs.
- * Output: JSON array with one result per file.
- */
-
-import { readFileSync } from "node:fs";
-import { outputJSON, stripCodeBlocks, stripInlineCode } from "./slop-helpers.ts";
-
-/** Resolve path relative to this script file. jiti provides __dirname at runtime. */
-function resolveAdjacent(filename: string): string {
-  const dir = (typeof __dirname !== "undefined" ? __dirname : "").replace(/\+$/, "");
-  return `${dir}/${filename}`;
-}
-
-interface VocabEntry {
-  term: string;
-  score: number;
-}
-
-interface VocabData {
-  tier1: VocabEntry[];
-  tier2: VocabEntry[];
-  tier3: VocabEntry[];
-  tier4: VocabEntry[];
-}
-
-const vocabulary: VocabData = JSON.parse(
-  readFileSync(resolveAdjacent("../references/vocabulary.json"), "utf-8"),
-);
-
-/** Build flat list with tier annotation. */
-function allVocab(): Array<VocabEntry & { tier: 1 | 2 | 3 | 4 }> {
-  return [
-    ...vocabulary.tier1.map((e) => ({ ...e, tier: 1 as const })),
-    ...vocabulary.tier2.map((e) => ({ ...e, tier: 2 as const })),
-    ...vocabulary.tier3.map((e) => ({ ...e, tier: 3 as const })),
-    ...vocabulary.tier4.map((e) => ({ ...e, tier: 4 as const })),
-  ];
-}
-
-interface VocabHit {
-  term: string;
-  tier: number;
-  score: number;
-  count: number;
-  context?: string;
-}
-
-interface VocabResult {
-  file: string;
-  wordCount: number;
-  totalScore: number;
-  normalizedScore: number;
-  tierScores: { tier1: number; tier2: number; tier3: number; tier4: number };
-  hits: VocabHit[];
-  rating: "clean" | "light" | "moderate" | "heavy";
-  recommendation: string;
-}
-
-function rate(normalizedScore: number): Pick<VocabResult, "rating" | "recommendation"> {
-  if (normalizedScore <= 1.0) {
-    return { rating: "clean", recommendation: "No action needed — vocabulary is clean." };
-  }
-  if (normalizedScore <= 2.5) {
-    return {
-      rating: "light",
-      recommendation: "Spot remediation — fix individual markers found above.",
-    };
-  }
-  if (normalizedScore <= 5.0) {
-    return {
-      rating: "moderate",
-      recommendation: "Section rewrite recommended — review flagged areas.",
-    };
-  }
-  return {
-    rating: "heavy",
-    recommendation: "Full document review — do not commit without addressing flagged markers.",
-  };
-}
-
-/** Escape special regex characters in a string. */
-function escapeRegex(s: string): string {
-  return s.replace(/[.*+?^${}()|[\]\\]/g, "\\$&");
-}
-
-/** Read a file as UTF-8 string. */
-function readFile(path: string): string {
-  return readFileSync(path, "utf-8");
-}
-
-/** Scan a single file for vocabulary markers. */
-function scanFile(filePath: string): VocabResult {
-  const content = readFile(filePath);
-  // Strip code blocks and inline code to avoid false positives from quoted terms
-  const prose = stripInlineCode(stripCodeBlocks(content));
-  const lowerContent = prose.toLowerCase();
-  const wordCount = prose.split(/[\s\n]+/).filter((w) => w.length > 0).length;
-
-  const hits: VocabHit[] = [];
-  let totalScore = 0;
-
-  for (const entry of allVocab()) {
-    const pattern = escapeRegex(entry.term.toLowerCase());
-    const re = new RegExp(pattern, "gi");
-    const matches = [...lowerContent.matchAll(re)];
-
-    if (matches.length === 0) continue;
-
-    const count = matches.length;
-    totalScore += count * entry.score;
-
-    // Context snippet around first occurrence (uses prose indices)
-    const idx = matches[0].index;
-    const start = Math.max(0, idx - 30);
-    const end = Math.min(prose.length, idx + entry.term.length + 30);
-    const context = prose.slice(start, end).replace(/\n/g, " ").trim();
-
-    hits.push({ term: entry.term, tier: entry.tier, score: entry.score, count, context });
-  }
-
-  hits.sort((a, b) => b.count * b.score - a.count * a.score);
-
-  const tierScores = {
-    tier1: hits.filter((h) => h.tier === 1).reduce((s, h) => s + h.count * h.score, 0),
-    tier2: hits.filter((h) => h.tier === 2).reduce((s, h) => s + h.count * h.score, 0),
-    tier3: hits.filter((h) => h.tier === 3).reduce((s, h) => s + h.count * h.score, 0),
-    tier4: hits.filter((h) => h.tier === 4).reduce((s, h) => s + h.count * h.score, 0),
-  };
-
-  const normalizedScore = wordCount > 0 ? (totalScore / wordCount) * 100 : 0;
-
-  return {
-    file: filePath,
-    wordCount,
-    totalScore,
-    normalizedScore: Math.round(normalizedScore * 100) / 100,
-    tierScores,
-    hits,
-    ...rate(normalizedScore),
-  };
-}
-
-// --- CLI ---
-const files = process.argv.slice(2);
-if (files.length === 0) {
-  process.stderr.write("Usage: pnpm exec jiti scripts/slop-scan-vocab.ts <file> [<file>...]\n");
-  process.exit(1);
-}
-
-const results = files.map(scanFile);
-outputJSON(results);

package/skills/supi-flow-slop-detect/scripts/slop-scan.ts

@@ -1,209 +0,0 @@
-#!/usr/bin/env -S pnpm exec jiti
-/**
- * Combined slop scanner — runs vocabulary + structural detection and produces
- * the final density score matching the formula in SKILL.md.
- *
- * Usage:
- *   pnpm exec jiti scripts/slop-scan.ts <file> [<file>...]
- *   pnpm exec jiti scripts/slop-scan.ts <file> --json-only  # raw JSON, no summary
- *
- * Cross-platform Node.js/TypeScript — runs wherever pi runs.
- * Output: Human-readable summary (default) or JSON (with --json-only).
- */
-
-import { execFileSync } from "node:child_process";
-import { type DocProfile, outputJSON } from "./slop-helpers.ts";
-
-interface VocabResult {
-  file: string;
-  wordCount: number;
-  totalScore: number;
-  normalizedScore: number;
-  tierScores: { tier1: number; tier2: number; tier3: number; tier4: number };
-  hits: Array<{ term: string; tier: number; score: number; count: number; context?: string }>;
-  rating: string;
-  recommendation: string;
-}
-
-interface StructuralResult {
-  file: string;
-  profile: DocProfile;
-  adjustments: string[];
-  wordCount: number;
-  metrics: {
-    emDashDensity: number;
-    bulletRatio: number;
-    participialTails: number;
-    participialTailsPer500: number;
-    arrowConnectors: number;
-    technicalArrowConnectors: number;
-    proseArrowConnectors: number;
-    correlativePairs: number;
-    plusSigns: number;
-    colons: number;
-    semicolons: number;
-    sentenceClusterRatio: number;
-    fromToRanges: number;
-    emojiBullets: number;
-    introBodyConclusion: boolean;
-    conclusionMirroring: boolean;
-    paragraphUniformity: number;
-  };
-  structuralScore: number;
-  flags: string[];
-}
-
-interface CombinedReport {
-  file: string;
-  profile: DocProfile;
-  adjustments: string[];
-  wordCount: number;
-  vocabScore: number;
-  structuralScore: number;
-  finalScore: number;
-  rating: "clean" | "light" | "moderate" | "heavy";
-  recommendation: string;
-  vocab: VocabResult;
-  structural: StructuralResult;
-}
-
-/** Get the directory of the currently running script. */
-function scriptDir(): string {
-  const dir = (typeof __dirname !== "undefined" ? __dirname : "").replace(/\\+$/, "");
-  return `${dir}/`;
-}
-
-function siblingPath(name: string): string {
-  return `${scriptDir()}${name}`;
-}
-
-function runScript(script: string, files: string[]): string {
-  return execFileSync("pnpm", ["exec", "jiti", script, ...files], {
-    encoding: "utf-8",
-    stdio: ["pipe", "pipe", "pipe"],
-  });
-}
-
-function rate(finalScore: number): { rating: CombinedReport["rating"]; recommendation: string } {
-  if (finalScore <= 1.0) return { rating: "clean", recommendation: "No action needed." };
-  if (finalScore <= 2.5)
-    return { rating: "light", recommendation: "Spot remediation — fix individual markers." };
-  if (finalScore <= 5.0)
-    return { rating: "moderate", recommendation: "Section rewrite recommended." };
-  return {
-    rating: "heavy",
-    recommendation: "Full document review — do not commit without fixing.",
-  };
-}
-
-function scanFile(filePath: string): CombinedReport {
-  const vocabOut = runScript(siblingPath("slop-scan-vocab.ts"), [filePath]);
-  const structOut = runScript(siblingPath("slop-scan-structural.ts"), [filePath]);
-
-  const vocabResults = JSON.parse(vocabOut) as VocabResult[];
-  const structResults = JSON.parse(structOut) as StructuralResult[];
-
-  const vocab = vocabResults[0];
-  const structural = structResults[0];
-  const finalScore = Math.min(vocab.normalizedScore + structural.structuralScore, 10);
-
-  return {
-    file: filePath,
-    profile: structural.profile,
-    adjustments: structural.adjustments,
-    wordCount: vocab.wordCount,
-    vocabScore: vocab.normalizedScore,
-    structuralScore: structural.structuralScore,
-    finalScore: Math.round(finalScore * 100) / 100,
-    vocab,
-    structural,
-    ...rate(finalScore),
-  };
-}
-
-// --- CLI ---
-const args = process.argv.slice(2);
-const jsonOnly = args.includes("--json-only");
-const files = args.filter((a) => a !== "--json-only");
-
-if (files.length === 0) {
-  console.error("Usage: pnpm exec jiti scripts/slop-scan.ts <file> [<file>...] [--json-only]");
-  process.exit(1);
-}
-
-const results = files.map(scanFile);
-
-if (jsonOnly) {
-  outputJSON(results);
-} else {
-  console.log("╔══════════════════════════════════════════════════╗");
-  console.log("║           Slop Detection Scan Report            ║");
-  console.log("╚══════════════════════════════════════════════════╝");
-  console.log();
-
-  for (const r of results) {
-    console.log(`File: ${r.file}`);
-    console.log(`Profile: ${r.profile}`);
-    console.log(`Words: ${r.wordCount}`);
-    if (r.adjustments.length > 0) {
-      console.log(`Adjustments: ${r.adjustments.join(", ")}`);
-    }
-    console.log("───");
-    console.log(`  Vocab score:       ${r.vocabScore.toFixed(2)}  (per 100 words)`);
-    console.log(`  Structural score:  ${r.structuralScore}  (penalty points)`);
-    console.log(`  Final score:       ${r.finalScore.toFixed(2)}`);
-
-    const barLen = Math.round(r.finalScore);
-    const bar = "█".repeat(Math.min(barLen, 10)) + "░".repeat(Math.max(0, 10 - barLen));
-    console.log(`  [${bar}]`);
-
-    let statusColor = "🟢";
-    if (r.rating === "light") statusColor = "🟡";
-    else if (r.rating === "moderate") statusColor = "🟠";
-    else if (r.rating === "heavy") statusColor = "🔴";
-
-    console.log(`  ${statusColor} ${r.rating.toUpperCase()} — ${r.recommendation}`);
-    console.log();
-
-    if (r.vocab.hits.length > 0) {
-      console.log("  Vocabulary markers found:");
-      for (const hit of r.vocab.hits.slice(0, 10)) {
-        console.log(
-          `    • Tier ${hit.tier} "${hit.term}" ×${hit.count} (score: ${hit.count * hit.score})`,
-        );
-      }
-      if (r.vocab.hits.length > 10) {
-        console.log(`    ... and ${r.vocab.hits.length - 10} more`);
-      }
-      console.log();
-    }
-
-    if (r.structural.flags.length > 0) {
-      console.log("  Structural flags:");
-      for (const flag of r.structural.flags) {
-        console.log(`    • ${flag}`);
-      }
-      console.log();
-    }
-
-    console.log("  Structural metrics:");
-    const m = r.structural.metrics;
-    console.log(`    Em dash density:        ${m.emDashDensity.toFixed(1)}/1000 words`);
-    console.log(`    Sentence clustering:    ${(m.sentenceClusterRatio * 100).toFixed(0)}%`);
-    console.log(`    Bullet ratio:           ${(m.bulletRatio * 100).toFixed(0)}%`);
-    console.log(`    Paragraph uniformity:   ${(m.paragraphUniformity * 100).toFixed(0)}%`);
-    console.log(`    Participial tails:      ${m.participialTailsPer500.toFixed(1)}/500 words`);
-    console.log(`    Correlative pairs:      ${m.correlativePairs}`);
-    console.log(`    Arrow connectors:       ${m.arrowConnectors}`);
-    console.log(`    Technical arrow chains: ${m.technicalArrowConnectors}`);
-    console.log(`    Prose arrow shorthand:  ${m.proseArrowConnectors}`);
-    console.log(`    Plus-sign conjunctions: ${m.plusSigns}`);
-    console.log(`    Emoji bullets:          ${m.emojiBullets}`);
-    console.log(`    Colons vs semicolons:   ${m.colons} / ${m.semicolons}`);
-    console.log(`    From→To ranges:         ${m.fromToRanges}`);
-    console.log(`    Intro-body-conclusion:  ${m.introBodyConclusion}`);
-    console.log(`    Conclusion mirroring:   ${m.conclusionMirroring}`);
-    console.log("───");
-    console.log();
-  }
-}