@tekyzinc/gsd-t 3.29.11 → 4.0.12

package/bin/context-budget-audit.cjs

@@ -1,447 +0,0 @@
-#!/usr/bin/env node
-// Context Budget Audit — measures the static context cost of a Claude Code session
-// before any user work happens. Reports tokens consumed by CLAUDE.md files, command
-// files, MCP server tool descriptions, and skills. Used to diagnose why long-running
-// sessions hit the manual compaction prompt.
-//
-// Usage:
-//   node bin/context-budget-audit.js                  # current project + global
-//   node bin/context-budget-audit.js --json           # JSON output for tooling
-//   node bin/context-budget-audit.js --top 20         # top N largest files
-//   node bin/context-budget-audit.js --threshold 5000 # flag files above N tokens
-//   node bin/context-budget-audit.js --model claude-opus-4-7  # size % to a model's window (default: 1M)
-
-const fs = require('fs');
-const path = require('path');
-const os = require('os');
-
-const { windowForModel, SAFE_DEFAULT_WINDOW } = require('./model-windows.cjs');
-
-// Token estimation: GPT/Claude tokenizers average ~4 chars/token for English+code.
-// This is a fast deterministic estimate, not a true tokenizer call. Within ~10%.
-const CHARS_PER_TOKEN = 4;
-
-// Context window for the "% of window" math. This is a static-analysis CLI with
-// no live transcript, so the model is resolved from --model / GSD_T_MODEL when
-// provided, else the model-windows safe LARGE default (1M). A bare 200K literal
-// (the old value) overstated every percentage 5× for Opus/Sonnet sessions.
-let CONTEXT_WINDOW = (() => {
-  const fromEnv = typeof process.env.GSD_T_MODEL === 'string' ? process.env.GSD_T_MODEL : '';
-  const argIdx = process.argv.indexOf('--model');
-  const fromArg = argIdx !== -1 ? process.argv[argIdx + 1] : '';
-  const modelId = fromArg || fromEnv;
-  return modelId ? windowForModel(modelId) : SAFE_DEFAULT_WINDOW;
-})();
-
-function estimateTokens(bytes) {
-  return Math.round(bytes / CHARS_PER_TOKEN);
-}
-
-function fmtPct(n) {
-  return `${n.toFixed(1)}%`;
-}
-
-function fmtNum(n) {
-  return n.toLocaleString('en-US');
-}
-
-function safeStat(p) {
-  try {
-    return fs.statSync(p);
-  } catch {
-    return null;
-  }
-}
-
-function safeRead(p) {
-  try {
-    return fs.readFileSync(p, 'utf8');
-  } catch {
-    return null;
-  }
-}
-
-function listFiles(dir, ext = null) {
-  if (!fs.existsSync(dir)) return [];
-  try {
-    return fs
-      .readdirSync(dir)
-      .filter((f) => !ext || f.endsWith(ext))
-      .map((f) => path.join(dir, f))
-      .filter((p) => {
-        const s = safeStat(p);
-        return s && s.isFile();
-      });
-  } catch {
-    return [];
-  }
-}
-
-function measureFile(filePath) {
-  const stat = safeStat(filePath);
-  if (!stat) return null;
-  return {
-    path: filePath,
-    bytes: stat.size,
-    tokens: estimateTokens(stat.size),
-  };
-}
-
-function measureCategory(name, files, source) {
-  const measured = files
-    .map(measureFile)
-    .filter(Boolean)
-    .sort((a, b) => b.tokens - a.tokens);
-  const totalBytes = measured.reduce((s, f) => s + f.bytes, 0);
-  const totalTokens = measured.reduce((s, f) => s + f.tokens, 0);
-  return {
-    name,
-    source,
-    fileCount: measured.length,
-    totalBytes,
-    totalTokens,
-    pctOfWindow: (totalTokens / CONTEXT_WINDOW) * 100,
-    files: measured,
-  };
-}
-
-// Estimate MCP tool descriptions cost — each deferred tool exposes a name + ~50-150
-// chars of description in the system prompt. We can't read them directly, but we
-// know the count from the list provided in conversation context (22 in this session).
-function estimateMcpToolsCost(toolCount = 22, avgCharsPerTool = 120) {
-  const bytes = toolCount * avgCharsPerTool;
-  return {
-    name: 'MCP deferred tool manifest (Figma + Gmail + Calendar)',
-    source: '~/.claude/settings.json mcpServers (estimated)',
-    fileCount: toolCount,
-    totalBytes: bytes,
-    totalTokens: estimateTokens(bytes),
-    pctOfWindow: (estimateTokens(bytes) / CONTEXT_WINDOW) * 100,
-    files: [],
-  };
-}
-
-// Estimate built-in tool schemas cost — Read, Edit, Write, Bash, Glob, Grep, etc.
-// These are loaded into every session. Approx based on their JSONSchema sizes.
-function estimateBuiltinToolsCost() {
-  // Top-of-prompt tools observed: Agent, Bash, Edit, Glob, Grep, Read, ScheduleWakeup,
-  // Skill, ToolSearch, Write. Each is ~200-2000 tokens in JSONSchema form.
-  const tools = {
-    Agent: 1800,
-    Bash: 2200,
-    Edit: 600,
-    Glob: 300,
-    Grep: 800,
-    Read: 700,
-    ScheduleWakeup: 900,
-    Skill: 400,
-    ToolSearch: 500,
-    Write: 400,
-  };
-  const totalTokens = Object.values(tools).reduce((s, t) => s + t, 0);
-  return {
-    name: 'Built-in tool schemas (top-of-prompt)',
-    source: 'Claude Code system prompt (estimated)',
-    fileCount: Object.keys(tools).length,
-    totalBytes: totalTokens * CHARS_PER_TOKEN,
-    totalTokens,
-    pctOfWindow: (totalTokens / CONTEXT_WINDOW) * 100,
-    files: [],
-  };
-}
-
-// Estimate Claude Code system prompt overhead — the harness instructions, env block,
-// system reminders, gitStatus, etc. Observed in the prompt header.
-function estimateSystemPromptCost() {
-  // The system prompt header (everything before "You are an interactive agent") plus
-  // the doing-tasks/tone-style/session-guidance sections is ~6000-8000 tokens.
-  const tokens = 7000;
-  return {
-    name: 'Claude Code system prompt (instructions, env, gitStatus)',
-    source: 'Claude Code harness',
-    fileCount: 1,
-    totalBytes: tokens * CHARS_PER_TOKEN,
-    totalTokens: tokens,
-    pctOfWindow: (tokens / CONTEXT_WINDOW) * 100,
-    files: [],
-  };
-}
-
-function audit({ projectDir, globalDir, top = 10, threshold = 0 }) {
-  const categories = [];
-
-  // 1. System prompt overhead
-  categories.push(estimateSystemPromptCost());
-
-  // 2. Built-in tool schemas
-  categories.push(estimateBuiltinToolsCost());
-
-  // 3. MCP deferred tool manifest
-  categories.push(estimateMcpToolsCost());
-
-  // 4. Global CLAUDE.md
-  const globalClaude = path.join(globalDir, 'CLAUDE.md');
-  if (fs.existsSync(globalClaude)) {
-    categories.push(measureCategory('Global ~/.claude/CLAUDE.md', [globalClaude], globalClaude));
-  }
-
-  // 5. Project CLAUDE.md
-  const projectClaude = path.join(projectDir, 'CLAUDE.md');
-  if (fs.existsSync(projectClaude)) {
-    categories.push(measureCategory('Project CLAUDE.md', [projectClaude], projectClaude));
-  }
-
-  // 6. Auto-memory (MEMORY.md + entries)
-  const memoryDir = path.join(
-    globalDir,
-    'projects',
-    `-${projectDir.replace(/\//g, '-').replace(/^-/, '')}`,
-    'memory'
-  );
-  if (fs.existsSync(memoryDir)) {
-    const memFiles = listFiles(memoryDir, '.md');
-    categories.push(measureCategory('Auto-memory (MEMORY.md + entries)', memFiles, memoryDir));
-  }
-
-  // 7. Installed user commands — IMPORTANT: Claude Code's skill system loads only
-  //    the manifest (name + first-line description) into the system prompt for each
-  //    command file. The full body loads only when the skill is invoked. So the
-  //    static cost is much smaller than the file size suggests — roughly ~150 chars
-  //    per command (name + description) regardless of body length.
-  const userCommandsDir = path.join(globalDir, 'commands');
-  if (fs.existsSync(userCommandsDir)) {
-    const cmdFiles = listFiles(userCommandsDir, '.md');
-    const manifestBytesPerCmd = 200; // typical "- name: <desc>" line
-    const manifestBytes = cmdFiles.length * manifestBytesPerCmd;
-    const fullBodyBytes = cmdFiles.reduce((s, f) => {
-      const st = safeStat(f);
-      return s + (st ? st.size : 0);
-    }, 0);
-    categories.push({
-      name: 'User commands MANIFEST (~/.claude/commands/ — names+descriptions only)',
-      source: userCommandsDir,
-      fileCount: cmdFiles.length,
-      totalBytes: manifestBytes,
-      totalTokens: estimateTokens(manifestBytes),
-      pctOfWindow: (estimateTokens(manifestBytes) / CONTEXT_WINDOW) * 100,
-      files: cmdFiles
-        .map((p) => {
-          const st = safeStat(p);
-          return st
-            ? { path: p, bytes: manifestBytesPerCmd, tokens: estimateTokens(manifestBytesPerCmd) }
-            : null;
-        })
-        .filter(Boolean),
-    });
-    // Also report what the FULL bodies cost when invoked, so trimming targets are visible
-    categories.push({
-      name: 'User command FULL BODIES (loaded only when each skill is invoked)',
-      source: userCommandsDir + ' (per-invocation cost, not baseline)',
-      fileCount: cmdFiles.length,
-      totalBytes: fullBodyBytes,
-      totalTokens: estimateTokens(fullBodyBytes),
-      pctOfWindow: (estimateTokens(fullBodyBytes) / CONTEXT_WINDOW) * 100,
-      files: cmdFiles.map(measureFile).filter(Boolean).sort((a, b) => b.tokens - a.tokens),
-      lazyLoaded: true,
-    });
-  }
-
-  // 8. Project commands — same lazy-load semantics as user commands. Only count if
-  //    it's a different directory AND the user's project actually exposes them as
-  //    skills (most projects don't).
-  const projectCommandsDir = path.join(projectDir, 'commands');
-  if (
-    fs.existsSync(projectCommandsDir) &&
-    path.resolve(projectCommandsDir) !== path.resolve(userCommandsDir)
-  ) {
-    const cmdFiles = listFiles(projectCommandsDir, '.md');
-    const manifestBytesPerCmd = 200;
-    const manifestBytes = cmdFiles.length * manifestBytesPerCmd;
-    const fullBodyBytes = cmdFiles.reduce((s, f) => {
-      const st = safeStat(f);
-      return s + (st ? st.size : 0);
-    }, 0);
-    categories.push({
-      name: 'Project commands MANIFEST (commands/ — names+descriptions only)',
-      source: projectCommandsDir,
-      fileCount: cmdFiles.length,
-      totalBytes: manifestBytes,
-      totalTokens: estimateTokens(manifestBytes),
-      pctOfWindow: (estimateTokens(manifestBytes) / CONTEXT_WINDOW) * 100,
-      files: [],
-    });
-    categories.push({
-      name: 'Project command FULL BODIES (per-invocation, not baseline)',
-      source: projectCommandsDir,
-      fileCount: cmdFiles.length,
-      totalBytes: fullBodyBytes,
-      totalTokens: estimateTokens(fullBodyBytes),
-      pctOfWindow: (estimateTokens(fullBodyBytes) / CONTEXT_WINDOW) * 100,
-      files: cmdFiles.map(measureFile).filter(Boolean).sort((a, b) => b.tokens - a.tokens),
-      lazyLoaded: true,
-    });
-  }
-
-  // 9. Project skills directory (if any)
-  const skillsDir = path.join(globalDir, 'skills');
-  if (fs.existsSync(skillsDir)) {
-    const skillFiles = listFiles(skillsDir, '.md');
-    if (skillFiles.length > 0) {
-      categories.push(measureCategory('Skills (~/.claude/skills/)', skillFiles, skillsDir));
-    }
-  }
-
-  // Compute totals — exclude lazy-loaded categories from the baseline cost
-  const baselineCats = categories.filter((c) => !c.lazyLoaded);
-  const totalTokens = baselineCats.reduce((s, c) => s + c.totalTokens, 0);
-  const totalBytes = baselineCats.reduce((s, c) => s + c.totalBytes, 0);
-  const totalPct = (totalTokens / CONTEXT_WINDOW) * 100;
-  const remaining = CONTEXT_WINDOW - totalTokens;
-
-  return {
-    contextWindow: CONTEXT_WINDOW,
-    totalBytes,
-    totalTokens,
-    totalPct,
-    remaining,
-    remainingPct: (remaining / CONTEXT_WINDOW) * 100,
-    categories,
-    top,
-    threshold,
-  };
-}
-
-function renderReport(result) {
-  const lines = [];
-  const bar = (pct, width = 40) => {
-    const filled = Math.round((pct / 100) * width);
-    return '█'.repeat(Math.min(filled, width)) + '░'.repeat(Math.max(width - filled, 0));
-  };
-
-  lines.push('');
-  lines.push('═══════════════════════════════════════════════════════════════════');
-  lines.push('  CONTEXT BUDGET AUDIT — what consumes context before you type');
-  lines.push('═══════════════════════════════════════════════════════════════════');
-  lines.push('');
-  lines.push(`  Context window:        ${fmtNum(result.contextWindow)} tokens`);
-  lines.push(`  Static preamble cost:  ${fmtNum(result.totalTokens)} tokens (${fmtPct(result.totalPct)})`);
-  lines.push(`  Remaining for work:    ${fmtNum(result.remaining)} tokens (${fmtPct(result.remainingPct)})`);
-  lines.push('');
-  lines.push(`  [${bar(result.totalPct)}] ${fmtPct(result.totalPct)}`);
-  lines.push('');
-
-  if (result.totalPct >= 70) {
-    lines.push('  🔴 CRITICAL: preamble already consumes >70% of context window.');
-    lines.push('     Long-running tasks will trigger manual /compact prompts.');
-  } else if (result.totalPct >= 50) {
-    lines.push('  ⚠️  WARNING: preamble consumes >50% of context window.');
-    lines.push('     Single-pass edits on large files may trigger compaction.');
-  } else if (result.totalPct >= 30) {
-    lines.push('  ⚡ ELEVATED: preamble consumes >30% of context window.');
-    lines.push('     Multi-step workflows may approach compaction threshold.');
-  } else {
-    lines.push('  ✅ HEALTHY: preamble consumes <30% of context window.');
-  }
-  lines.push('');
-  lines.push('───────────────────────────────────────────────────────────────────');
-  lines.push('  BREAKDOWN BY CATEGORY (largest first)');
-  lines.push('───────────────────────────────────────────────────────────────────');
-  lines.push('');
-
-  const sortedCats = [...result.categories].sort((a, b) => b.totalTokens - a.totalTokens);
-  for (const cat of sortedCats) {
-    const tag = cat.lazyLoaded ? ' [LAZY — not in baseline]' : '';
-    lines.push(`  ${cat.name}${tag}`);
-    lines.push(`    source: ${cat.source}`);
-    lines.push(
-      `    files: ${cat.fileCount}   bytes: ${fmtNum(cat.totalBytes)}   tokens: ${fmtNum(cat.totalTokens)} (${fmtPct(cat.pctOfWindow)})`
-    );
-    lines.push(`    [${bar(cat.pctOfWindow, 30)}]`);
-    lines.push('');
-  }
-
-  // Top N files across all categories
-  const allFiles = result.categories
-    .flatMap((c) => c.files)
-    .sort((a, b) => b.tokens - a.tokens);
-  if (allFiles.length > 0) {
-    lines.push('───────────────────────────────────────────────────────────────────');
-    lines.push(`  TOP ${result.top} HEAVIEST FILES`);
-    lines.push('───────────────────────────────────────────────────────────────────');
-    lines.push('');
-    const topFiles = allFiles.slice(0, result.top);
-    for (const f of topFiles) {
-      const pct = (f.tokens / CONTEXT_WINDOW) * 100;
-      const flag = pct >= 5 ? ' 🔥' : pct >= 2 ? ' ⚠️' : '';
-      const rel = f.path.replace(os.homedir(), '~');
-      lines.push(`  ${fmtNum(f.tokens).padStart(7)} tok  ${fmtPct(pct).padStart(6)}  ${rel}${flag}`);
-    }
-    lines.push('');
-  }
-
-  // Files above threshold
-  if (result.threshold > 0) {
-    const above = allFiles.filter((f) => f.tokens >= result.threshold);
-    if (above.length > 0) {
-      lines.push('───────────────────────────────────────────────────────────────────');
-      lines.push(`  FILES ABOVE THRESHOLD (>= ${fmtNum(result.threshold)} tokens)`);
-      lines.push('───────────────────────────────────────────────────────────────────');
-      lines.push('');
-      for (const f of above) {
-        const rel = f.path.replace(os.homedir(), '~');
-        lines.push(`  ${fmtNum(f.tokens).padStart(7)} tok  ${rel}`);
-      }
-      lines.push('');
-    }
-  }
-
-  lines.push('───────────────────────────────────────────────────────────────────');
-  lines.push('  NOTES');
-  lines.push('───────────────────────────────────────────────────────────────────');
-  lines.push('');
-  lines.push('  - Token estimate uses 4 chars/token (within ~10% of real tokenizer).');
-  lines.push('  - "Static preamble" = everything loaded BEFORE you type your first');
-  lines.push('    message. Add ~10-20K tokens of typical conversation overhead.');
-  lines.push('  - Skills auto-load into the prompt as a list of names+descriptions.');
-  lines.push('    Their full bodies are loaded only when invoked, but the manifest');
-  lines.push('    itself contributes to baseline.');
-  lines.push('  - Built-in and MCP tool costs are estimates; real values vary by');
-  lines.push('    Claude Code version. Check ENABLE_TELEMETRY for exact counts.');
-  lines.push('');
-
-  return lines.join('\n');
-}
-
-function main() {
-  const args = process.argv.slice(2);
-  const opts = {
-    projectDir: process.cwd(),
-    globalDir: path.join(os.homedir(), '.claude'),
-    top: 10,
-    threshold: 0,
-    json: false,
-  };
-  for (let i = 0; i < args.length; i++) {
-    const a = args[i];
-    if (a === '--json') opts.json = true;
-    else if (a === '--top') opts.top = parseInt(args[++i], 10);
-    else if (a === '--threshold') opts.threshold = parseInt(args[++i], 10);
-    else if (a === '--project') opts.projectDir = path.resolve(args[++i]);
-    else if (a === '--global') opts.globalDir = path.resolve(args[++i]);
-    else if (a === '--model') i++; // consumed at init for window sizing; skip value
-    else if (a === '--help' || a === '-h') {
-      console.log('Usage: node bin/context-budget-audit.js [--json] [--top N] [--threshold N] [--model <claude-model-id>]');
-      process.exit(0);
-    }
-  }
-
-  const result = audit(opts);
-
-  if (opts.json) {
-    console.log(JSON.stringify(result, null, 2));
-  } else {
-    console.log(renderReport(result));
-  }
-}
-
-main();

package/bin/context-meter-config.cjs

@@ -1,101 +0,0 @@
-/**
- * Context Meter config loader (M34).
- *
- * Reads .gsd-t/context-meter-config.json, merges over defaults, validates,
- * and returns the resolved config. Missing file → defaults. Unknown schema
- * version or API-key leak → throws with a clear message.
- *
- * See .gsd-t/contracts/context-meter-contract.md for the schema, validation
- * rules, and the API-key-never-stored invariant.
- */
-
-const fs = require("fs");
-const path = require("path");
-
-const DEFAULTS = Object.freeze({
-  version: 1,
-  thresholdPct: 75,
-  modelWindowSize: 200000,
-  checkFrequency: 5,
-  apiKeyEnvVar: "ANTHROPIC_API_KEY",
-  statePath: ".gsd-t/.context-meter-state.json",
-  logPath: ".gsd-t/context-meter.log",
-  timeoutMs: 2000,
-});
-
-const SUPPORTED_VERSION = 1;
-const API_KEY_FIELD_RE = /api.?key/i;
-const HEX_LOOKALIKE_RE = /^[a-zA-Z0-9_-]{64,}$/;
-
-function loadConfig(projectRoot) {
-  const root = projectRoot || process.cwd();
-  const configPath = path.join(root, ".gsd-t", "context-meter-config.json");
-
-  let userConfig = {};
-  if (fs.existsSync(configPath)) {
-    const raw = fs.readFileSync(configPath, "utf8");
-    try {
-      userConfig = JSON.parse(raw);
-    } catch (e) {
-      throw new Error(
-        `context-meter-config: invalid JSON in ${configPath}: ${e.message}`
-      );
-    }
-    if (!userConfig || typeof userConfig !== "object" || Array.isArray(userConfig)) {
-      throw new Error(`context-meter-config: ${configPath} must contain a JSON object`);
-    }
-  }
-
-  validateNoKeyLeak(userConfig);
-
-  if (userConfig.version !== undefined && userConfig.version !== SUPPORTED_VERSION) {
-    throw new Error(
-      `context-meter-config: unsupported schema version ${userConfig.version} ` +
-        `(expected ${SUPPORTED_VERSION}). See .gsd-t/contracts/context-meter-contract.md#breaking-changes for migration.`
-    );
-  }
-
-  const merged = { ...DEFAULTS, ...userConfig, version: SUPPORTED_VERSION };
-  validateRanges(merged);
-  return merged;
-}
-
-function validateNoKeyLeak(obj) {
-  for (const key of Object.keys(obj)) {
-    if (key === "apiKeyEnvVar") continue;
-    if (API_KEY_FIELD_RE.test(key)) {
-      throw new Error(
-        `context-meter-config: field "${key}" looks like an API key storage field. ` +
-          `API keys must only be read from the env var named in apiKeyEnvVar.`
-      );
-    }
-    const val = obj[key];
-    if (typeof val === "string" && val.length > 100 && HEX_LOOKALIKE_RE.test(val)) {
-      throw new Error(
-        `context-meter-config: field "${key}" contains a long token-like string. ` +
-          `Do not store API keys in config — use apiKeyEnvVar to name the env var.`
-      );
-    }
-  }
-}
-
-function validateRanges(c) {
-  const assert = (cond, msg) => { if (!cond) throw new Error(`context-meter-config: ${msg}`); };
-
-  assert(Number.isFinite(c.thresholdPct) && c.thresholdPct > 0 && c.thresholdPct < 100,
-    `thresholdPct must be a number in (0, 100), got ${c.thresholdPct}`);
-  assert(Number.isInteger(c.modelWindowSize) && c.modelWindowSize > 0,
-    `modelWindowSize must be a positive integer, got ${c.modelWindowSize}`);
-  assert(Number.isInteger(c.checkFrequency) && c.checkFrequency >= 1,
-    `checkFrequency must be an integer >= 1, got ${c.checkFrequency}`);
-  assert(typeof c.apiKeyEnvVar === "string" && c.apiKeyEnvVar.length > 0,
-    `apiKeyEnvVar must be a non-empty string, got ${JSON.stringify(c.apiKeyEnvVar)}`);
-  assert(typeof c.statePath === "string" && c.statePath.length > 0,
-    `statePath must be a non-empty string`);
-  assert(typeof c.logPath === "string" && c.logPath.length > 0,
-    `logPath must be a non-empty string`);
-  assert(Number.isInteger(c.timeoutMs) && c.timeoutMs > 0,
-    `timeoutMs must be a positive integer, got ${c.timeoutMs}`);
-}
-
-module.exports = { loadConfig, DEFAULTS };

package/bin/context-meter-config.test.cjs

@@ -1,101 +0,0 @@
-const { test } = require("node:test");
-const assert = require("node:assert/strict");
-const fs = require("fs");
-const path = require("path");
-const os = require("os");
-
-const { loadConfig, DEFAULTS } = require("./context-meter-config.cjs");
-
-function makeProject(config) {
-  const root = fs.mkdtempSync(path.join(os.tmpdir(), "cmcfg-"));
-  fs.mkdirSync(path.join(root, ".gsd-t"), { recursive: true });
-  if (config !== undefined) {
-    fs.writeFileSync(
-      path.join(root, ".gsd-t", "context-meter-config.json"),
-      typeof config === "string" ? config : JSON.stringify(config)
-    );
-  }
-  return root;
-}
-
-test("missing config file → returns defaults", () => {
-  const root = makeProject();
-  assert.deepEqual(loadConfig(root), DEFAULTS);
-});
-
-test("valid full config → returns exact values", () => {
-  const custom = {
-    version: 1,
-    thresholdPct: 80,
-    modelWindowSize: 400000,
-    checkFrequency: 10,
-    apiKeyEnvVar: "CLAUDE_KEY",
-    statePath: ".gsd-t/my-state.json",
-    logPath: ".gsd-t/my.log",
-    timeoutMs: 5000,
-  };
-  const root = makeProject(custom);
-  assert.deepEqual(loadConfig(root), custom);
-});
-
-test("partial config → missing fields filled with defaults", () => {
-  const root = makeProject({ thresholdPct: 60, timeoutMs: 1000 });
-  const cfg = loadConfig(root);
-  assert.equal(cfg.thresholdPct, 60);
-  assert.equal(cfg.timeoutMs, 1000);
-  assert.equal(cfg.modelWindowSize, DEFAULTS.modelWindowSize);
-  assert.equal(cfg.checkFrequency, DEFAULTS.checkFrequency);
-  assert.equal(cfg.apiKeyEnvVar, DEFAULTS.apiKeyEnvVar);
-});
-
-test("thresholdPct out of range throws", () => {
-  for (const bad of [0, 100, -5, 150, "80"]) {
-    const root = makeProject({ thresholdPct: bad });
-    assert.throws(() => loadConfig(root), /thresholdPct/);
-  }
-});
-
-test("modelWindowSize <= 0 throws", () => {
-  for (const bad of [0, -1, 1.5, "100"]) {
-    const root = makeProject({ modelWindowSize: bad });
-    assert.throws(() => loadConfig(root), /modelWindowSize/);
-  }
-});
-
-test("checkFrequency < 1 throws", () => {
-  for (const bad of [0, -1, 0.5]) {
-    const root = makeProject({ checkFrequency: bad });
-    assert.throws(() => loadConfig(root), /checkFrequency/);
-  }
-});
-
-test("empty apiKeyEnvVar throws", () => {
-  const root = makeProject({ apiKeyEnvVar: "" });
-  assert.throws(() => loadConfig(root), /apiKeyEnvVar/);
-});
-
-test("unknown version throws with migration pointer", () => {
-  const root = makeProject({ version: 2, thresholdPct: 75 });
-  assert.throws(() => loadConfig(root), /version 2|migration/i);
-});
-
-test("config containing an apiKey field is rejected as leak", () => {
-  const root = makeProject({ apiKey: "sk-ant-abc123" });
-  assert.throws(() => loadConfig(root), /api.?key/i);
-});
-
-test("config containing a long hex-like string value is rejected as leak", () => {
-  const longHex = "a".repeat(120);
-  const root = makeProject({ customField: longHex });
-  assert.throws(() => loadConfig(root), /api.?key|token-like/i);
-});
-
-test("invalid JSON throws with clear message", () => {
-  const root = makeProject("{ not valid json");
-  assert.throws(() => loadConfig(root), /invalid JSON/);
-});
-
-test("non-object JSON throws", () => {
-  const root = makeProject("[1, 2, 3]");
-  assert.throws(() => loadConfig(root), /JSON object/);
-});