blitzstrike 2.4.121 → 2.4.122

package/CHANGELOG.md

@@ -5,6 +5,19 @@ All notable changes to Blitz Strike are documented here.
 The format follows [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/).
 
+## [2.4.122] — 2026-09-29
+
+### Added — `param_surface` (autonomous parameter discovery + specialist hints)
+
+- New `param_surface(target, class?)`: fetch a target and extract EVERY testable
+  parameter — URL query params, HTML form fields, path-segment identifiers
+  (/users/123 + UUIDs), and JS-referenced endpoints — then attach per-class
+  discovery + test hints so the agent tests the DISCOVERED params, not just the
+  hinted one. New `intelligence/specialist-discovery.json` maps 15 vuln classes
+  (xss/sqli/ssrf/idor/lfi/rce/xxe/jwt/open_redirect/prototype_pollution/csti/
+  mass_assignment/header_injection/api_security/file_upload) to where to look +
+  which tool to drive. (154 MCP tools.)
+
 ## [2.4.121] — 2026-09-29
 
 ### Added — empirical ledger auto-feed + benchmark coverage 30/30

package/dist/index.js

@@ -17452,10 +17452,10 @@ var init_variants = __esm(() => {
 });
 
 // src/sync.ts
-import { mkdirSync as mkdirSync10, writeFileSync as writeFileSync9, readFileSync as readFileSync30, readdirSync as readdirSync9, statSync as statSync6, rmSync as rmSync5 } from "node:fs";
-import { join as join29 } from "node:path";
+import { mkdirSync as mkdirSync10, writeFileSync as writeFileSync9, readFileSync as readFileSync31, readdirSync as readdirSync9, statSync as statSync6, rmSync as rmSync5 } from "node:fs";
+import { join as join30 } from "node:path";
 import { homedir as homedir11 } from "node:os";
-import { fileURLToPath as fileURLToPath17 } from "node:url";
+import { fileURLToPath as fileURLToPath18 } from "node:url";
 import { spawnSync as spawnSync4 } from "node:child_process";
 function dirExists(p) {
   try {
@@ -17470,7 +17470,7 @@ function countFiles(dir) {
     const walk = (d) => {
       for (const e of readdirSync9(d, { withFileTypes: true })) {
         if (e.isDirectory())
-          walk(join29(d, e.name));
+          walk(join30(d, e.name));
         else
           n++;
       }
@@ -17484,8 +17484,8 @@ function countFiles(dir) {
 function syncData() {
   if (!dirExists(DATA_ROOT2))
     mkdirSync10(DATA_ROOT2, { recursive: true });
-  const payloadsPath = join29(DATA_ROOT2, "payloads");
-  const templatesPath = join29(DATA_ROOT2, "templates");
+  const payloadsPath = join30(DATA_ROOT2, "payloads");
+  const templatesPath = join30(DATA_ROOT2, "templates");
   const hasPayloads = dirExists(payloadsPath) && countFiles(payloadsPath) > 0;
   const hasTemplates = dirExists(templatesPath) && countFiles(templatesPath) > 0;
   if (hasPayloads && hasTemplates) {
@@ -17495,8 +17495,8 @@ function syncData() {
       data_root: DATA_ROOT2
     };
   }
-  const localPayloads = join29(ROOT10, "payloads");
-  const localTemplates = join29(ROOT10, "templates");
+  const localPayloads = join30(ROOT11, "payloads");
+  const localTemplates = join30(ROOT11, "templates");
   let payloadsSynced = false;
   let templatesSynced = false;
   if (!hasPayloads && dirExists(localPayloads)) {
@@ -17509,7 +17509,7 @@ function syncData() {
   }
   const gitOk = spawnSync4("git", ["--version"]).status === 0;
   if (gitOk && (!dirExists(payloadsPath) || countFiles(payloadsPath) === 0 || !dirExists(templatesPath) || countFiles(templatesPath) === 0)) {
-    const tmp = join29(DATA_ROOT2, ".tmp-clone");
+    const tmp = join30(DATA_ROOT2, ".tmp-clone");
     if (dirExists(tmp))
       rmSync5(tmp, { recursive: true, force: true });
     const r = spawnSync4("git", ["clone", "--depth", "1", "--filter=blob:none", "--sparse", GITHUB_REPO, tmp], { timeout: 120000 });
@@ -17517,11 +17517,11 @@ function syncData() {
       const sparse = spawnSync4("git", ["-C", tmp, "sparse-checkout", "set", "payloads", "templates"], { timeout: 120000 });
       if (sparse.status === 0) {
         if (!dirExists(payloadsPath) || countFiles(payloadsPath) === 0) {
-          copyDir(join29(tmp, "payloads"), payloadsPath);
+          copyDir(join30(tmp, "payloads"), payloadsPath);
           payloadsSynced = true;
         }
         if (!dirExists(templatesPath) || countFiles(templatesPath) === 0) {
-          copyDir(join29(tmp, "templates"), templatesPath);
+          copyDir(join30(tmp, "templates"), templatesPath);
           templatesSynced = true;
         }
       }
@@ -17543,18 +17543,18 @@ function copyDir(src, dst) {
 function manualCopy(src, dst) {
   mkdirSync10(dst, { recursive: true });
   for (const e of readdirSync9(src, { withFileTypes: true })) {
-    const s = join29(src, e.name);
-    const d = join29(dst, e.name);
+    const s = join30(src, e.name);
+    const d = join30(dst, e.name);
     if (e.isDirectory())
       manualCopy(s, d);
     else
-      writeFileSync9(d, readFileSync30(s));
+      writeFileSync9(d, readFileSync31(s));
   }
 }
-var ROOT10, DATA_ROOT2, GITHUB_REPO = "https://github.com/shinthink/blitzstrike.git";
+var ROOT11, DATA_ROOT2, GITHUB_REPO = "https://github.com/shinthink/blitzstrike.git";
 var init_sync = __esm(() => {
-  ROOT10 = join29(fileURLToPath17(new URL(".", import.meta.url)), "..");
-  DATA_ROOT2 = process.env.BLITZSTRIKE_DATA ?? join29(homedir11(), ".blitzstrike", "data");
+  ROOT11 = join30(fileURLToPath18(new URL(".", import.meta.url)), "..");
+  DATA_ROOT2 = process.env.BLITZSTRIKE_DATA ?? join30(homedir11(), ".blitzstrike", "data");
 });
 
 // src/update.ts
@@ -17599,15 +17599,15 @@ var init_update = __esm(() => {
 });
 
 // src/release.ts
-import { readFileSync as readFileSync32, existsSync as existsSync17 } from "node:fs";
-import { join as join30 } from "node:path";
-import { fileURLToPath as fileURLToPath18 } from "node:url";
+import { readFileSync as readFileSync33, existsSync as existsSync17 } from "node:fs";
+import { join as join31 } from "node:path";
+import { fileURLToPath as fileURLToPath19 } from "node:url";
 import { execSync as execSync4 } from "node:child_process";
 function validateRelease() {
   const checks = [];
   checks.push({ name: "package.json version", ok: true, detail: VERSION });
   try {
-    const changelog = readFileSync32(join30(ROOT11, "CHANGELOG.md"), "utf8");
+    const changelog = readFileSync33(join31(ROOT12, "CHANGELOG.md"), "utf8");
     const m = changelog.match(/^##\s+\[([0-9]+\.[0-9]+\.[0-9]+)\]/m);
     const changelogVersion = m?.[1] ?? null;
     checks.push({
@@ -17619,7 +17619,7 @@ function validateRelease() {
     checks.push({ name: "CHANGELOG version", ok: false, detail: "CHANGELOG.md not found" });
   }
   try {
-    const tag = execSync4("git describe --tags --abbrev=0", { cwd: ROOT11, encoding: "utf8", stdio: ["ignore", "pipe", "ignore"] }).trim();
+    const tag = execSync4("git describe --tags --abbrev=0", { cwd: ROOT12, encoding: "utf8", stdio: ["ignore", "pipe", "ignore"] }).trim();
     checks.push({ name: "Git tag", ok: tag === `v${VERSION}`, detail: tag, soft: true });
   } catch {
     checks.push({ name: "Git tag", ok: true, detail: "unavailable (no git/tags)", soft: true });
@@ -17639,22 +17639,22 @@ function validateRelease() {
   return { version: VERSION, checks, consistent: checks.filter((c) => !c.soft).every((c) => c.ok) };
 }
 function lockfilePresent() {
-  return existsSync17(join30(ROOT11, "package-lock.json")) || existsSync17(join30(ROOT11, "bun.lock")) || existsSync17(join30(ROOT11, "bun.lockb"));
+  return existsSync17(join31(ROOT12, "package-lock.json")) || existsSync17(join31(ROOT12, "bun.lock")) || existsSync17(join31(ROOT12, "bun.lockb"));
 }
-var ROOT11;
+var ROOT12;
 var init_release = __esm(() => {
   init_version();
-  ROOT11 = join30(fileURLToPath18(new URL(".", import.meta.url)), "..");
+  ROOT12 = join31(fileURLToPath19(new URL(".", import.meta.url)), "..");
 });
 
 // src/dependency.ts
-import { readFileSync as readFileSync33 } from "node:fs";
-import { join as join31 } from "node:path";
-import { fileURLToPath as fileURLToPath19 } from "node:url";
+import { readFileSync as readFileSync34 } from "node:fs";
+import { join as join32 } from "node:path";
+import { fileURLToPath as fileURLToPath20 } from "node:url";
 function checkDependencies() {
   let pkg = {};
   try {
-    pkg = JSON.parse(readFileSync33(join31(ROOT12, "package.json"), "utf8"));
+    pkg = JSON.parse(readFileSync34(join32(ROOT13, "package.json"), "utf8"));
   } catch {
     return { install_scripts: [], git_dependencies: [], total_dependencies: 0, safe: false };
   }
@@ -17675,16 +17675,16 @@ function checkDependencies() {
     safe: installScripts.length === 0 && gitDependencies.length === 0
   };
 }
-var ROOT12;
+var ROOT13;
 var init_dependency = __esm(() => {
-  ROOT12 = join31(fileURLToPath19(new URL(".", import.meta.url)), "..");
+  ROOT13 = join32(fileURLToPath20(new URL(".", import.meta.url)), "..");
 });
 
 // src/cli.ts
-import { readFileSync as readFileSync34, existsSync as existsSync18, writeFileSync as writeFileSync10, mkdirSync as mkdirSync11 } from "node:fs";
-import { join as join32, dirname as dirname6 } from "node:path";
+import { readFileSync as readFileSync35, existsSync as existsSync18, writeFileSync as writeFileSync10, mkdirSync as mkdirSync11 } from "node:fs";
+import { join as join33, dirname as dirname6 } from "node:path";
 import { homedir as homedir12 } from "node:os";
-import { fileURLToPath as fileURLToPath20 } from "node:url";
+import { fileURLToPath as fileURLToPath21 } from "node:url";
 import { spawnSync as spawnSync5 } from "node:child_process";
 function which(cmd) {
   return spawnSync5("sh", ["-c", 'command -v "$1" >/dev/null 2>&1', "sh", cmd]).status === 0;
@@ -17727,7 +17727,7 @@ async function runDoctor() {
     detail: h1 ? `X-HackerOne-Research: ${h1}` : "H1_USERNAME not set",
     fix: h1 ? undefined : "export H1_USERNAME=<your-hackerone-username> (adds X-HackerOne-Research to outbound requests)"
   });
-  const dataOk = existsSync18(join32(ROOT13, "chains.json")) && existsSync18(join32(ROOT13, "tools-catalog.json"));
+  const dataOk = existsSync18(join33(ROOT14, "chains.json")) && existsSync18(join33(ROOT14, "tools-catalog.json"));
   issues.push({
     name: "Data layers (chains + tools-catalog)",
     status: dataOk ? "ok" : "fail",
@@ -17778,21 +17778,21 @@ BlitzStrike — ${verdict}`);
 function resolveCommand() {
   if (which("blitzstrike"))
     return { command: "blitzstrike", args: ["serve", "--mcp"] };
-  const src = join32(ROOT13, "src", "index.ts");
+  const src = join33(ROOT14, "src", "index.ts");
   if (which("bun") && existsSync18(src))
     return { command: "bun", args: ["run", src, "serve", "--mcp"] };
   if (which("npx"))
     return { command: "npx", args: ["-y", "blitzstrike", "serve", "--mcp"] };
   if (which("bunx"))
     return { command: "bunx", args: ["blitzstrike", "serve", "--mcp"] };
-  const dist = join32(ROOT13, "dist", "index.js");
+  const dist = join33(ROOT14, "dist", "index.js");
   if (existsSync18(dist))
     return { command: "bun", args: [dist, "serve", "--mcp"] };
   return { command: "npx", args: ["-y", "blitzstrike", "serve", "--mcp"] };
 }
 function readJson(p) {
   try {
-    return JSON.parse(readFileSync34(p, "utf8"));
+    return JSON.parse(readFileSync35(p, "utf8"));
   } catch {
     return null;
   }
@@ -17825,17 +17825,17 @@ function opencodeWrite(p) {
   };
 }
 function copyOpenCodeAgents() {
-  const srcDir = join32(ROOT13, "opencode-agents");
-  const dstDir = join32(homedir12(), ".config", "opencode", "agents");
+  const srcDir = join33(ROOT14, "opencode-agents");
+  const dstDir = join33(homedir12(), ".config", "opencode", "agents");
   if (!existsSync18(srcDir))
     return;
   if (!existsSync18(dstDir))
     mkdirSync11(dstDir, { recursive: true });
   for (const f of ["Blitz Strike.md", "Blitz.md", "Eagle Eye.md", "Strike.md"]) {
-    const src = join32(srcDir, f);
-    const dst = join32(dstDir, f);
+    const src = join33(srcDir, f);
+    const dst = join33(dstDir, f);
     if (existsSync18(src)) {
-      writeFileSync10(dst, readFileSync34(src, "utf8"));
+      writeFileSync10(dst, readFileSync35(src, "utf8"));
     }
   }
 }
@@ -17844,7 +17844,7 @@ function codexWrite(p) {
     const dir = dirname6(p);
     if (!existsSync18(dir))
       mkdirSync11(dir, { recursive: true });
-    let existing = existsSync18(p) ? readFileSync34(p, "utf8") : "";
+    let existing = existsSync18(p) ? readFileSync35(p, "utf8") : "";
     if (!existing.trimEnd().endsWith(`
 `))
       existing += `
@@ -17863,7 +17863,7 @@ function hermesWrite(p) {
     const dir = dirname6(p);
     if (!existsSync18(dir))
       mkdirSync11(dir, { recursive: true });
-    let existing = existsSync18(p) ? readFileSync34(p, "utf8") : "";
+    let existing = existsSync18(p) ? readFileSync35(p, "utf8") : "";
     existing = existing.replace(/^  blitzstrike:\n(?:    .*\n?)*/m, "");
     if (!existing.trimEnd().endsWith(`
 `))
@@ -17890,74 +17890,74 @@ mcp_servers:
 function detectAgents() {
   const home = homedir12();
   const agents = [];
-  const claudeJson = join32(home, ".claude.json");
+  const claudeJson = join33(home, ".claude.json");
   agents.push({
     name: "Claude Code",
     detectPath: claudeJson,
     installed: () => existsSync18(claudeJson) || which("claude"),
     write: jsonMcpServersWrite(claudeJson)
   });
-  const claudeDesktop = join32(home, ".config", "Claude", "claude_desktop_config.json");
+  const claudeDesktop = join33(home, ".config", "Claude", "claude_desktop_config.json");
   agents.push({
     name: "Claude Desktop",
     detectPath: claudeDesktop,
     installed: () => existsSync18(claudeDesktop),
     write: jsonMcpServersWrite(claudeDesktop)
   });
-  const cursor = join32(home, ".cursor", "mcp.json");
+  const cursor = join33(home, ".cursor", "mcp.json");
   agents.push({
     name: "Cursor",
     detectPath: cursor,
-    installed: () => existsSync18(join32(home, ".cursor")) || which("cursor"),
+    installed: () => existsSync18(join33(home, ".cursor")) || which("cursor"),
     write: jsonMcpServersWrite(cursor)
   });
-  const opencode = join32(home, ".config", "opencode", "opencode.jsonc");
+  const opencode = join33(home, ".config", "opencode", "opencode.jsonc");
   agents.push({
     name: "OpenCode",
     detectPath: opencode,
-    installed: () => existsSync18(join32(home, ".config", "opencode")) || which("opencode"),
+    installed: () => existsSync18(join33(home, ".config", "opencode")) || which("opencode"),
     write: opencodeWrite(opencode)
   });
-  const codex = join32(home, ".codex", "config.toml");
+  const codex = join33(home, ".codex", "config.toml");
   agents.push({
     name: "Codex",
     detectPath: codex,
-    installed: () => existsSync18(join32(home, ".codex")) || which("codex"),
+    installed: () => existsSync18(join33(home, ".codex")) || which("codex"),
     write: codexWrite(codex)
   });
-  const hermes = join32(home, ".hermes", "config.yaml");
+  const hermes = join33(home, ".hermes", "config.yaml");
   agents.push({
     name: "Hermes",
     detectPath: hermes,
-    installed: () => existsSync18(join32(home, ".hermes")) || which("hermes"),
+    installed: () => existsSync18(join33(home, ".hermes")) || which("hermes"),
     write: hermesWrite(hermes)
   });
-  const gemini = join32(home, ".gemini", "settings.json");
+  const gemini = join33(home, ".gemini", "settings.json");
   agents.push({
     name: "Gemini CLI",
     detectPath: gemini,
-    installed: () => existsSync18(join32(home, ".gemini")) || which("gemini"),
+    installed: () => existsSync18(join33(home, ".gemini")) || which("gemini"),
     write: jsonMcpServersWrite(gemini)
   });
-  const windsurf = join32(home, ".codeium", "windsurf", "mcp_config.json");
+  const windsurf = join33(home, ".codeium", "windsurf", "mcp_config.json");
   agents.push({
     name: "Windsurf",
     detectPath: windsurf,
-    installed: () => existsSync18(join32(home, ".codeium")) || which("windsurf"),
+    installed: () => existsSync18(join33(home, ".codeium")) || which("windsurf"),
     write: jsonMcpServersWrite(windsurf)
   });
-  const copilot = join32(home, ".copilot", "mcp.json");
+  const copilot = join33(home, ".copilot", "mcp.json");
   agents.push({
     name: "Copilot",
     detectPath: copilot,
-    installed: () => existsSync18(join32(home, ".copilot")),
+    installed: () => existsSync18(join33(home, ".copilot")),
     write: jsonMcpServersWrite(copilot)
   });
-  const cline = join32(home, ".cline", "mcp_settings.json");
+  const cline = join33(home, ".cline", "mcp_settings.json");
   agents.push({
     name: "Cline",
     detectPath: cline,
-    installed: () => existsSync18(join32(home, ".cline")),
+    installed: () => existsSync18(join33(home, ".cline")),
     write: jsonMcpServersWrite(cline)
   });
   return agents;
@@ -18069,13 +18069,13 @@ Failed (${failedList.length}) — reason per tool:`);
 Common causes: missing toolchain, pip/git network failures, or very heavy builds. ` + `Re-run after installing the missing toolchain — failed tools are skipped, not fatal.`);
   }
 }
-var ROOT13;
+var ROOT14;
 var init_cli = __esm(() => {
   init_catalog();
   init_release();
   init_dependency();
   init_mcp_status();
-  ROOT13 = join32(fileURLToPath20(new URL(".", import.meta.url)), "..");
+  ROOT14 = join33(fileURLToPath21(new URL(".", import.meta.url)), "..");
 });
 
 // node_modules/zod/v3/helpers/util.js
@@ -32713,7 +32713,7 @@ class StdioServerTransport {
 
 // src/server.ts
 init_scanner();
-import { readFileSync as readFileSync31 } from "node:fs";
+import { readFileSync as readFileSync32 } from "node:fs";
 import { execFileSync as execFileSync8 } from "node:child_process";
 
 // src/review-harness.ts
@@ -60508,19 +60508,118 @@ async function nucleiScan(opts) {
   };
 }
 
-// src/benchmark.ts
-import { readFileSync as readFileSync25, readdirSync as readdirSync7, writeFileSync as writeFileSync6, rmSync as rmSync3, mkdirSync as mkdirSync7 } from "node:fs";
+// src/param-surface.ts
+import { readFileSync as readFileSync25 } from "node:fs";
 import { join as join25 } from "node:path";
 import { fileURLToPath as fileURLToPath15 } from "node:url";
+var ROOT8 = join25(fileURLToPath15(new URL(".", import.meta.url)), "..");
+function loadSpecialists() {
+  try {
+    return JSON.parse(readFileSync25(join25(ROOT8, "intelligence", "specialist-discovery.json"), "utf8")).classes;
+  } catch {
+    return [];
+  }
+}
+function formFields(html) {
+  const out = new Set;
+  const re = /<(?:input|select|textarea|button)\b[^>]*\bname\s*=\s*["']([^"']+)["']/gi;
+  let m;
+  while ((m = re.exec(html)) !== null)
+    if (m[1] && !/^(csrf|_csrf|token|_token)$/i.test(m[1]))
+      out.add(m[1]);
+  return [...out];
+}
+function pathParams(urls) {
+  const seen = new Set;
+  const out = [];
+  const uuidRe = /[0-9a-f]{8}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{12}/i;
+  for (const u of urls) {
+    let path = "";
+    try {
+      path = new URL(u).pathname;
+    } catch {
+      continue;
+    }
+    const segs = path.split("/").filter(Boolean);
+    for (let i = 0;i < segs.length; i++) {
+      const s = segs[i];
+      const prev = segs[i - 1] ?? "id";
+      if (/^\d+$/.test(s)) {
+        const name = prev.replace(/[^a-z0-9_]/gi, "").toLowerCase() || "id";
+        const key = `${name}:id`;
+        if (!seen.has(key)) {
+          seen.add(key);
+          out.push({ name, sample: s, kind: "id" });
+        }
+      } else if (uuidRe.test(s)) {
+        const name = prev.replace(/[^a-z0-9_]/gi, "").toLowerCase() || "uuid";
+        const key = `${name}:uuid`;
+        if (!seen.has(key)) {
+          seen.add(key);
+          out.push({ name, sample: s, kind: "uuid" });
+        }
+      }
+    }
+  }
+  return out;
+}
+function jsRefs(html) {
+  const out = new Set;
+  for (const m of html.matchAll(/(?:fetch|axios(?:\.get|\.post)?|XMLHttpRequest\.open)\s*\(\s*["'`]([^"'`]+)["'`]/gi)) {
+    const u = m[1];
+    if (/^https?:\/\//i.test(u) || u.startsWith("/"))
+      out.add(u);
+  }
+  return [...out].slice(0, 50);
+}
+async function paramSurface(target, class_) {
+  const base = target.replace(/\/$/, "");
+  let html = "";
+  let status = 0;
+  try {
+    const res = await fetch(target, { redirect: "follow", headers: { "User-Agent": "Mozilla/5.0 (compatible; BlitzStrike/2.4)" } });
+    status = res.status;
+    html = await res.text();
+  } catch {}
+  const { links, scripts } = crawlLinks(base, html);
+  const allUrls = [target, ...links, ...scripts];
+  const query = extractParams(allUrls);
+  const fields = formFields(html);
+  const paths = pathParams(allUrls);
+  const js = jsRefs(html);
+  const result = {
+    target,
+    url: base,
+    query_params: query,
+    form_fields: fields,
+    path_params: paths,
+    js_refs: js,
+    total: query.length + fields.length + paths.length
+  };
+  if (class_) {
+    const spec = loadSpecialists();
+    const c = class_.toLowerCase().trim();
+    const matched = spec.filter((s) => s.id === c || s.id.includes(c) || c.includes(s.id));
+    result.specialist_hints = matched.length > 0 ? matched : spec.filter((s) => s.id.includes(c.slice(0, 3)));
+  } else {
+    result.specialist_hints = loadSpecialists();
+  }
+  return { ...result, url: base, status_hint: status === 0 ? "fetch failed" : `http ${status}` };
+}
+
+// src/benchmark.ts
+import { readFileSync as readFileSync26, readdirSync as readdirSync7, writeFileSync as writeFileSync6, rmSync as rmSync3, mkdirSync as mkdirSync7 } from "node:fs";
+import { join as join26 } from "node:path";
+import { fileURLToPath as fileURLToPath16 } from "node:url";
 import { tmpdir as tmpdir3, homedir as homedir9 } from "node:os";
 init_complex_bugs();
-var ROOT8 = join25(fileURLToPath15(new URL(".", import.meta.url)), "..");
-var CORPUS = join25(ROOT8, "bench", "corpus");
-var CAPTURED_CORPUS_DIR = process.env.BLITZSTRIKE_CORPUS_DIR ?? join25(homedir9(), ".blitzstrike", "corpus");
+var ROOT9 = join26(fileURLToPath16(new URL(".", import.meta.url)), "..");
+var CORPUS = join26(ROOT9, "bench", "corpus");
+var CAPTURED_CORPUS_DIR = process.env.BLITZSTRIKE_CORPUS_DIR ?? join26(homedir9(), ".blitzstrike", "corpus");
 function parseCorpusFile(path) {
   let text;
   try {
-    text = readFileSync25(path, "utf8");
+    text = readFileSync26(path, "utf8");
   } catch {
     return null;
   }
@@ -60576,7 +60675,7 @@ function loadDir(dir) {
     for (const f of readdirSync7(dir)) {
       if (!f.endsWith(".md"))
         continue;
-      const e = parseCorpusFile(join25(dir, f));
+      const e = parseCorpusFile(join26(dir, f));
       if (e)
         entries.push(e);
     }
@@ -60603,7 +60702,7 @@ function captureFalsePositive(input) {
     input.code,
     "```"
   ].filter((l) => l !== null);
-  const path = join25(CAPTURED_CORPUS_DIR, `${id}.md`);
+  const path = join26(CAPTURED_CORPUS_DIR, `${id}.md`);
   writeFileSync6(path, lines.join(`
 `) + `
 `, "utf8");
@@ -60625,7 +60724,7 @@ function detect(entry) {
   }
   const lang = entry.language;
   if (lang === "php") {
-    const tmp = join25(tmpdir3(), `blitzstrike-bench-${entry.id}-${process.pid}.php`);
+    const tmp = join26(tmpdir3(), `blitzstrike-bench-${entry.id}-${process.pid}.php`);
     writeFileSync6(tmp, entry.code);
     const r = analyzeDataFlow2(tmp);
     rmSync3(tmp, { force: true });
@@ -60694,12 +60793,12 @@ function runBenchmark() {
 }
 
 // src/obligations.ts
-import { readFileSync as readFileSync26, appendFileSync as appendFileSync4, mkdirSync as mkdirSync8, existsSync as existsSync13, writeFileSync as writeFileSync7 } from "node:fs";
-import { join as join26 } from "node:path";
+import { readFileSync as readFileSync27, appendFileSync as appendFileSync4, mkdirSync as mkdirSync8, existsSync as existsSync13, writeFileSync as writeFileSync7 } from "node:fs";
+import { join as join27 } from "node:path";
 import { createHash as createHash6 } from "node:crypto";
 import { homedir as homedir10 } from "node:os";
-var HOME_DIR4 = process.env.BLITZSTRIKE_HOME ?? join26(homedir10(), ".blitzstrike");
-var OBLIGATION_PATH = join26(HOME_DIR4, "obligations.jsonl");
+var HOME_DIR4 = process.env.BLITZSTRIKE_HOME ?? join27(homedir10(), ".blitzstrike");
+var OBLIGATION_PATH = join27(HOME_DIR4, "obligations.jsonl");
 function ensureDir3() {
   if (!existsSync13(HOME_DIR4))
     mkdirSync8(HOME_DIR4, { recursive: true });
@@ -60712,7 +60811,7 @@ function loadObligations() {
     return [];
   try {
     const out = [];
-    for (const line of readFileSync26(OBLIGATION_PATH, "utf8").split(`
+    for (const line of readFileSync27(OBLIGATION_PATH, "utf8").split(`
 `)) {
       const t = line.trim();
       if (!t)
@@ -60816,11 +60915,11 @@ init_framework_security();
 // src/variant-scan.ts
 init_scanner();
 init_complex_bugs();
-import { readFileSync as readFileSync28 } from "node:fs";
+import { readFileSync as readFileSync29 } from "node:fs";
 
 // src/cross-file.ts
 init_scanner();
-import { readFileSync as readFileSync27 } from "node:fs";
+import { readFileSync as readFileSync28 } from "node:fs";
 var PUBLIC_HOOK_RE = /(?:wp_ajax_nopriv_\w+|admin_post_nopriv_\w+|init|template_redirect|wp_loaded|parse_request|user_register|um_user_register|register_new_user|authenticate)/;
 var PRIV_SINK_RE = /\b(set_role|add_role|remove_role|set_user_role|wp_set_current_user|wp_set_auth_cookie|wp_insert_user|wp_create_user|wp_update_user)\s*\(/;
 var ROLE_META_RE = /update_user_meta\s*\(\s*[^,]+,\s*['"](wp_capabilities|role|roles|capabilities|wp_user_level)['"]/;
@@ -60872,7 +60971,7 @@ function detectCrossFilePrivesc(root) {
   for (const file of files) {
     let code;
     try {
-      code = stripComments(readFileSync27(file, "utf8"));
+      code = stripComments(readFileSync28(file, "utf8"));
     } catch {
       continue;
     }
@@ -60913,7 +61012,7 @@ function detectCrossFilePrivesc(root) {
   for (const file of files) {
     let code;
     try {
-      code = stripComments(readFileSync27(file, "utf8"));
+      code = stripComments(readFileSync28(file, "utf8"));
     } catch {
       continue;
     }
@@ -60936,7 +61035,7 @@ function detectCrossFilePrivesc(root) {
   for (const file of files) {
     let code;
     try {
-      code = stripComments(readFileSync27(file, "utf8"));
+      code = stripComments(readFileSync28(file, "utf8"));
     } catch {
       continue;
     }
@@ -61051,7 +61150,7 @@ function variantScan(root, opts = {}) {
   for (const file of files) {
     let code;
     try {
-      code = readFileSync28(file, "utf8");
+      code = readFileSync29(file, "utf8");
     } catch {
       continue;
     }
@@ -62736,7 +62835,7 @@ function listPocTypes() {
 import { execSync as execSync3 } from "node:child_process";
 import { mkdtempSync as mkdtempSync3, mkdirSync as mkdirSync9, writeFileSync as writeFileSync8, rmSync as rmSync4, existsSync as existsSync15 } from "node:fs";
 import { tmpdir as tmpdir4 } from "node:os";
-import { join as join27 } from "node:path";
+import { join as join28 } from "node:path";
 function findForge() {
   try {
     const p = execSync3("which forge 2>/dev/null", { encoding: "utf8", timeout: 3000 }).trim();
@@ -62785,14 +62884,14 @@ async function foundryRun(input) {
     };
   }
   const poc = input.poc_code ?? foundryPoc(input.class_ ?? "missing_access_control", name);
-  const dir = mkdtempSync3(join27(tmpdir4(), "blitzstrike-foundry-"));
+  const dir = mkdtempSync3(join28(tmpdir4(), "blitzstrike-foundry-"));
   try {
     try {
       execSync3(`${forge} init --force --no-git ${dir}`, { stdio: "ignore", timeout: 60000 });
     } catch {
-      mkdirSync9(join27(dir, "src"), { recursive: true });
-      mkdirSync9(join27(dir, "test"), { recursive: true });
-      writeFileSync8(join27(dir, "foundry.toml"), `[profile.default]
+      mkdirSync9(join28(dir, "src"), { recursive: true });
+      mkdirSync9(join28(dir, "test"), { recursive: true });
+      writeFileSync8(join28(dir, "foundry.toml"), `[profile.default]
 src = "src"
 out = "out"
 libs = ["lib"]
@@ -62802,12 +62901,12 @@ solc = "0.8.20"
     for (const f of ["Counter.sol", "Counter.t.sol", "Counter.s.sol"]) {
       for (const sub of ["src", "test", "script"]) {
         try {
-          rmSync4(join27(dir, sub, f), { force: true });
+          rmSync4(join28(dir, sub, f), { force: true });
         } catch {}
       }
     }
-    writeFileSync8(join27(dir, "src", `${name}.sol`), input.contract_code);
-    writeFileSync8(join27(dir, "test", `${name}.t.sol`), poc);
+    writeFileSync8(join28(dir, "src", `${name}.sol`), input.contract_code);
+    writeFileSync8(join28(dir, "test", `${name}.t.sol`), poc);
     const { out } = runForge(forge, dir, input.timeout_ms ?? 60000);
     const tests = parseTestResults(out);
     const anyFail = tests.some((t) => t.status === "fail");
@@ -62825,18 +62924,18 @@ solc = "0.8.20"
 }
 
 // src/enterprise-benchmark.ts
-import { existsSync as existsSync16, readFileSync as readFileSync29, readdirSync as readdirSync8, statSync as statSync5 } from "node:fs";
-import { join as join28, resolve } from "node:path";
-import { fileURLToPath as fileURLToPath16 } from "node:url";
+import { existsSync as existsSync16, readFileSync as readFileSync30, readdirSync as readdirSync8, statSync as statSync5 } from "node:fs";
+import { join as join29, resolve } from "node:path";
+import { fileURLToPath as fileURLToPath17 } from "node:url";
 init_complex_bugs();
-var ROOT9 = resolve(fileURLToPath16(new URL(".", import.meta.url)), "..", "bench", "enterprise");
+var ROOT10 = resolve(fileURLToPath17(new URL(".", import.meta.url)), "..", "bench", "enterprise");
 var ENTERPRISE_FIXTURES = [
-  { id: "php-laravel", dir: join28(ROOT9, "php-laravel"), expected_vulns: 4 },
-  { id: "js-express", dir: join28(ROOT9, "js-express"), expected_vulns: 2 },
-  { id: "py-django", dir: join28(ROOT9, "py-django"), expected_vulns: 2 },
-  { id: "java-spring", dir: join28(ROOT9, "java-spring"), expected_vulns: 2 },
-  { id: "wp-batch-route", dir: join28(ROOT9, "wp-batch-route"), expected_vulns: 3 },
-  { id: "php-complex", dir: join28(ROOT9, "php-complex"), expected_vulns: 4 }
+  { id: "php-laravel", dir: join29(ROOT10, "php-laravel"), expected_vulns: 4 },
+  { id: "js-express", dir: join29(ROOT10, "js-express"), expected_vulns: 2 },
+  { id: "py-django", dir: join29(ROOT10, "py-django"), expected_vulns: 2 },
+  { id: "java-spring", dir: join29(ROOT10, "java-spring"), expected_vulns: 2 },
+  { id: "wp-batch-route", dir: join29(ROOT10, "wp-batch-route"), expected_vulns: 3 },
+  { id: "php-complex", dir: join29(ROOT10, "php-complex"), expected_vulns: 4 }
 ];
 function sourceFiles(root) {
   const out = [];
@@ -62848,7 +62947,7 @@ function sourceFiles(root) {
       return;
     }
     for (const e of entries) {
-      const p = join28(dir, e);
+      const p = join29(dir, e);
       let st;
       try {
         st = statSync5(p);
@@ -62869,7 +62968,7 @@ function detectFixture(dir) {
   for (const f of sourceFiles(dir)) {
     let code;
     try {
-      code = readFileSync29(f, "utf8");
+      code = readFileSync30(f, "utf8");
     } catch {
       continue;
     }
@@ -63082,7 +63181,7 @@ function createServer2() {
 ` + `- A hard block (Cloudflare/WAF/IP-block/captcha) is NOT the end: mark the finding blocked + note the required vantage (residential / authorized / correct-SNI / --resolve direct-to-origin), then CONTINUE to the next lead (origin bypass, non-prod cluster, mobile APK, subdomain).
 ` + `- Keep trying every lead AND every chain step until each is verified/executed or blocked, THEN generate_report. Never deliver a bare surface map, and never stop at 'found X' when X enables more.
 
-` + "TWO PATHS:\\n" + "- Source path on disk -> blitz_scan / taint_file / eagle_eye2 (static).\\n" + "- Live URL -> live_recon / active_scan / strike_verify (web-hunting).\\n\\n" + "DEEPEN A FINDING: attack_plan(target, tech, params) to get the prioritized vector plan (decide what to test next with data); list_chains (escalation), chain_links (compose), technique_lookup(cls), payload_lookup, bypass_lookup (blocked). Ground any detector hit empirically with pattern_lookup(id) — it returns the class's CWE, real-world signature, escalation chains, citable references, per-class bypass techniques, and the high-value assets (crown jewels) that class hits, so a raw hit becomes a class-aware finding. Resolve free-form terminology first with synonym_lookup(term) (IDOR/BOLA/SQLi/prompt-injection → canonical id). For an llm_injection hit, its pattern carries a validation discipline — the RUN-TWICE RULE (reproduce token-for-token across two fresh sessions), NON-GUESSABLE ANCHOR (must leak a real key/URL, not a plausible guess), and OOB PROOF (a verifiable callback log) — apply all three before reporting. BEFORE generating a report, run submission_gate(...) and only ship a finding that passes all 7 questions (reproducible, in-scope, real-impact, no-privileged-assumption, verified-live, evidence-redacted, severity-derived); one false = kill it. Generate the concrete reproduction recipe with auto_poc(type, base_url, endpoint, param) — it emits the exact marker request + negative control + expected differential + Patchstack-style title, so 'reproducible' is executable, not abstract. Calibrate severity with severity_calibrate(type, cvss) to get a P1–P5 tier + VRT-style category. Start each engagement with engagement_scaffold(target, scope) for the folder + scope.md structure, and check data_refresh() to confirm the data layers are consistent. Phase doctrine: read_skill(bs-orchestrate-engagement | bs-web-hunting | bs-source-audit | bs-verify-finding). Campaign plan: read_playbook(target-type). On any failure call retry_guidance(signal) / model_fallback(signal) before giving up. HUNTING DOCTRINE (maximize payout rate): (1) SIBLING RULE — if one endpoint requires auth, check every sibling (/export /delete /share); ~30% of paid IDOR/auth bugs come from siblings — AUTOMATE it with sibling_scan(base_url, endpoint, token_a, token_b), the deterministic live logic-bug prober (sibling enumeration + differential IDOR + auth matrix + mass assignment + rate limit, evidence-first). (2) A→B SIGNAL — a confirmed bug signals a class of developer mistake; hunt B and C (20-min box) before reporting. (3) TWO-ACCOUNT IDOR — always test with attacker account A reading victim account B, and state it in the report. (4) PoC ESCALATION — IDOR shows victim data (not 200), XSS shows cookie exfil (not alert), SSRF shows internal response (not DNS), SQLi shows DB content (not error). (5) FOLLOW THE MONEY — billing/credits/refunds/wallet flows. (6) Kill weak findings with kill_list(type) BEFORE the gate — missing headers, self-XSS, open-redirect-alone, SSRF-DNS-only, GraphQL-introspection-alone are always N/A standalone. (7) CI/CD + SAML/SSO are critical attack surfaces (pull_request_target, expression injection, XML signature wrapping, comment injection). Full doctrine: hunting_doctrine(). WHAT TO HUNT NEXT: watchlist() returns the ranked CWE frequency + high-value CVEs with payout + a saturated flag (prioritize high-freq/high-payout/not-saturated: missing-authz, SQLi, SSRF, deserialization; skip saturated XSS/CSRF/open-redirect unless they chain into takeover). The CVE watchlist now spans 4,000+ product fingerprints (live from nuclei-templates when installed — refresh with sync_cve_corpus() before a big recon pass); live_recon's suggested_cves names the SPECIFIC CVE for the detected product/version — then cve_verify(cve, target) CONFIRMS it live by running the matching nuclei template (VERIFIED only when nuclei actually matches; a CVE is report-ready only after cve_verify, never from the fingerprint alone). For 'scan this target' automation, nuclei_scan(target) fingerprints the stack and runs the WHOLE matched template set (CVE templates + optional broad tech-tag) in one rate-limited pass — use it instead of verifying CVEs one-by-one. SURFACE-FIRST: surface_map(surface) maps an attack surface (image/redirect proxy, edge cache, payment portal, AI agent, file service, Electron, SSO) to its highest-value tests — map surfaces then test, don't spray payloads. PRECISION GUARD: resolve_check(host) detects when an ISP/country redirects the domain to a blocking/sinkhole page (gambling/adult/streaming niches) — audit the PUBLIC IP, never the sinkhole.\n" + "KNOWLEDGE BASE (discover, don't guess): skillLookup(topic) / listSkills for the 40 attack-path + hunting skills (ap-*, hp-*, bs-*); list_attack_vectors / attack_vectors(cat) for the 34-category vector taxonomy; taxonomy(kind) for OWASP/CWE/ASVS/API-top-10; read_playbook(type) for the 17 campaign playbooks; list_manuals / read_tool_manual for 317 tool references; list_tools / run_catalog_tool for the 141-tool catalog. Every tool, skill, and data layer is reachable from these. CROSS-TARGET INTELLIGENCE: strike_resolve auto-records confirmed/false_positive verdicts into the empirical ledger; query it with intelligence(action=hit_rates/top_vectors/recall) so a new target on a known tech inherits historical hit-rates (attack_plan already blends them into success_probability). intelligence(action=recall, tech, target) is the per-engagement pattern recall — it returns which patterns historically CONFIRMED on that tech (prioritize them) versus which only ever produced false positives (avoid them), so a new engagement starts from the empirical playbook of past ones. LIVE PENTEST + TARGETING (black-box, WAF-safe): cache_gap(base_url, endpoint) fuzzes path-normalization + static-extension variants for cache deception/poisoning; blind_oracle(base_url, param, value) detects which SINK a param reaches (sql/ssti/xss) via benign arithmetic (id=1+0, test${7*7}) WITHOUT a malicious payload; param_precedence(base_url, param, A, B) maps duplicate-parameter precedence (first/last-wins) for WAF-bypass; broken_link_scan(url) detects expired-domain takeover (crawls outbound refs → DNS/RDAP → flags claimable domains). Logic-bug live probers: jwt_forge(token, endpoint, …) forges JWT confusion tokens (alg:none / RS256→HS256 key-confusion / kid-jku-x5u) and proves which is ACCEPTED live; proto_pollution_probe(target, endpoint) sends __proto__ / constructor.prototype JSON canaries and reports VERIFIED only on reflection; smuggling_test(target) sends CL.TE / TE.CL desync probes over a raw connection. OOB (out-of-band) PROOF for BLIND vulns: oob_start() spins an interactsh listener (unique *.oast.* domain, streams DNS/HTTP interactions — the standard for public targets) or a local HTTP listener for local labs; inject the canary, then oob_poll(id) until the hit arrives; oob_stop(id) to clean up. A blind SSRF/XXE/SQLi/command-injection is CONFIRMED only when the canary actually arrives — never from a guess. SOURCE TARGETING: review_harness(path) FIRST for full coverage — it applies a fixed 5-signal rubric (entry_points/auth_gates/sinks/secrets/logic_bugs) to EVERY source file and sorts the signal-dense ones first (a complete map, not a partial scan); then action_surface(code) maps every wp_ajax_* action to its vuln class + DERIVED priority (what to look for WHERE); security_state(sanitizers, sink) gives a wrong-context sanitization verdict; framework_knowledge(framework) returns auth/nonce primitives for 9 frameworks; variant_scan(root, signature) mass-sweeps an install base for ONE bug family; patch_analyze(old, new) REVERSES a security patch into the 0-day, and its mine_signature feeds variant_scan to sweep unpatched copies. All of these are DETECTION HINTS — verify each live (strike_verify / browser_validate) before it enters the report. NEWEST DETECTORS — understand AND try them: variant_scan also emits priv_esc (privilege escalation — a role/capability/user change fed by request input, traced cross-file via reverse call-graph, e.g. set_role($id, sanitize_key($_POST['um-role'])) or update_user_meta(...,'wp_capabilities', $_POST[...])) and flags uploadAllow=>array('all') unrestricted-upload configs as file_upload (the file-manager/connector shape). TO TRY a priv_esc hit: submit the role field with 'administrator' (or send wp_capabilities meta) and confirm the account actually gains admin. TO TRY an unrestricted-upload hit: POST a .php/.phtml and confirm it lands web-reachable + executes. variant_scan also emits hardcoded_secret (CWE-798) in three tiers: a secret field (api_key/db_password/access_token/aws_*) assigned a literal (HIGH), a known secret FORMAT (JWT eyJ…/AWS AKIA…/GitHub ghp_…/Slack xoxb-…/…) regardless of name (HIGH), or a high-entropy suspect token (MEDIUM — verify before reporting). Verify a frontend/JS secret live via drive_devtools (inspect the JS bundle / network response and confirm the key is actually exposed), and for a backend secret confirm it is reachable/valid. A static hit is a HYPOTHESIS until verified live — never report unverified. For a repo, also scan_git_history(root) — secrets DELETED from HEAD still live in .git history (the removed-later-is-not-fixed trap), and it feeds every recovered pre-deletion blob through the same secret detector. Once you recover a credential, prove it is still ACTIVE with validate_leaked_key(apiBase, endpoints, key|blobUrl) — a READ-ONLY differential test (401-without-key vs 200-with-key = auth bypass PROVEN). Two supply-chain detectors: dependency_confusion (a manifest/build file resolving a package from a public registry — pypi.org/npmjs.org --extra-index-url, or a scoped @company/pkg with no private registry) and ml_supply_chain (torch.load without weights_only=True / joblib.load / np.load(allow_pickle=True) / keras load_model on an untrusted model → pickle RCE). Rust is supported end-to-end: the taint engine (blitz_scan / taint_file) maps Rust sinks (std::process::Command, sqlx/diesel, reqwest/hyper, std::fs, serde/bincode) to the classic classes, plus two Rust-specific detectors — rust_unsafe (transmute / assume_init / from_raw_parts / ptr::* — memory-unsafety surface, CWE-119) and rust_format_string (println!/format! with a variable as the format string — CWE-134). The sqlx query! compile-time macro is treated as safe (suppressed). Two more web-class detectors: jwt_alg_confusion (a JWT verifier that dispatches on the token's alg header and accepts BOTH RS256 and HS256, with the HS256 HMAC secret derived from public/shared key material — CWE-347; the attacker sets alg=HS256 and signs with the public key to forge admin tokens) and cache_deception (a reverse proxy caching static-extension URLs with a cache key that omits the session cookie, over a session-authenticated backend — CWE-525; the victim's cached private response is re-fetched anonymously). TO TRY a jwt_alg_confusion hit: fetch the public key (/.well-known/jwks.json), forge an HS256 token signed with that public key, and confirm admin access. TO TRY a cache_deception hit: request an authenticated path with a static extension appended (.css/.js) and re-fetch it anonymously. Three more detectors: cors_misconfiguration (Access-Control-Allow-Origin reflects the request Origin or is '*' while credentials are true — CWE-942; any site reads the victim's authenticated responses), xss (user input into a DOM sink — innerHTML / dangerouslySetInnerHTML / document.write / .html() / eval — CWE-79), and subdomain_takeover (a dangling CNAME/AliasTarget to a claimable service like *.github.io/*.herokuapp.com/*.amazonaws.com — CWE-404). TO TRY a cors hit: send an evil Origin + confirm the response echoes it with Allow-Credentials:true. TO TRY an xss hit: confirm the sink executes (drive_devtools). TO TRY a takeover hit: verify the CNAME target resolves to an unclaimed/not-found page. Two more client-side detectors: postmessage_no_origin_check (a window message listener that reads event.data but never validates event.origin — CWE-345; any site can drive the handler) and client_trusted_flag (a privilege/entitlement flag like isAdmin/withVip/premium read from request input and used in a security decision — CWE-807/639; derive the flag from server state, never the client). TO TRY a postMessage hit: send a crafted message from an attacker origin and confirm the handler acts on it. TO TRY a client-trusted-flag hit: flip the flag param to true and confirm the privilege is actually granted. WEB3: for Solidity/EVM, call web3_audit(code|path) — a deterministic 13-class smart-contract audit (reentrancy incl. ERC721/ERC1155 callback reentrancy, unchecked return, unchecked arithmetic, tx.origin auth, signature replay, unprotected selfdestruct, delegatecall-to-user-input, missing access control, timestamp dependence, spot-price oracle manipulation incl. slot0/latestRoundData, unbounded loop, abi.encodePacked hash collision, missing address(0) validation) returning ranked findings (class/CWE/severity/line), then foundry_poc(class, contract) to generate an executable Foundry test, and foundry_run(contract_code, poc_code|class_) to actually EXECUTE it — a PASSING PoC test is the proof that the exploit reproduces (deterministic: forge's own test verdict). A web3 hit is a HYPOTHESIS until the Foundry PoC reproduces it.\n" + `PRECISION ANALYSIS — chrome-devtools MCP is MANDATORY, not optional:
+` + "TWO PATHS:\\n" + "- Source path on disk -> blitz_scan / taint_file / eagle_eye2 (static).\\n" + "- Live URL -> live_recon / active_scan / strike_verify (web-hunting).\\n\\n" + "DEEPEN A FINDING: attack_plan(target, tech, params) to get the prioritized vector plan (decide what to test next with data); list_chains (escalation), chain_links (compose), technique_lookup(cls), payload_lookup, bypass_lookup (blocked). Ground any detector hit empirically with pattern_lookup(id) — it returns the class's CWE, real-world signature, escalation chains, citable references, per-class bypass techniques, and the high-value assets (crown jewels) that class hits, so a raw hit becomes a class-aware finding. Resolve free-form terminology first with synonym_lookup(term) (IDOR/BOLA/SQLi/prompt-injection → canonical id). For an llm_injection hit, its pattern carries a validation discipline — the RUN-TWICE RULE (reproduce token-for-token across two fresh sessions), NON-GUESSABLE ANCHOR (must leak a real key/URL, not a plausible guess), and OOB PROOF (a verifiable callback log) — apply all three before reporting. BEFORE generating a report, run submission_gate(...) and only ship a finding that passes all 7 questions (reproducible, in-scope, real-impact, no-privileged-assumption, verified-live, evidence-redacted, severity-derived); one false = kill it. Generate the concrete reproduction recipe with auto_poc(type, base_url, endpoint, param) — it emits the exact marker request + negative control + expected differential + Patchstack-style title, so 'reproducible' is executable, not abstract. Calibrate severity with severity_calibrate(type, cvss) to get a P1–P5 tier + VRT-style category. Start each engagement with engagement_scaffold(target, scope) for the folder + scope.md structure, and check data_refresh() to confirm the data layers are consistent. Phase doctrine: read_skill(bs-orchestrate-engagement | bs-web-hunting | bs-source-audit | bs-verify-finding). Campaign plan: read_playbook(target-type). On any failure call retry_guidance(signal) / model_fallback(signal) before giving up. HUNTING DOCTRINE (maximize payout rate): (1) SIBLING RULE — if one endpoint requires auth, check every sibling (/export /delete /share); ~30% of paid IDOR/auth bugs come from siblings — AUTOMATE it with sibling_scan(base_url, endpoint, token_a, token_b), the deterministic live logic-bug prober (sibling enumeration + differential IDOR + auth matrix + mass assignment + rate limit, evidence-first). (2) A→B SIGNAL — a confirmed bug signals a class of developer mistake; hunt B and C (20-min box) before reporting. (3) TWO-ACCOUNT IDOR — always test with attacker account A reading victim account B, and state it in the report. (4) PoC ESCALATION — IDOR shows victim data (not 200), XSS shows cookie exfil (not alert), SSRF shows internal response (not DNS), SQLi shows DB content (not error). (5) FOLLOW THE MONEY — billing/credits/refunds/wallet flows. (6) Kill weak findings with kill_list(type) BEFORE the gate — missing headers, self-XSS, open-redirect-alone, SSRF-DNS-only, GraphQL-introspection-alone are always N/A standalone. (7) CI/CD + SAML/SSO are critical attack surfaces (pull_request_target, expression injection, XML signature wrapping, comment injection). Full doctrine: hunting_doctrine(). WHAT TO HUNT NEXT: watchlist() returns the ranked CWE frequency + high-value CVEs with payout + a saturated flag (prioritize high-freq/high-payout/not-saturated: missing-authz, SQLi, SSRF, deserialization; skip saturated XSS/CSRF/open-redirect unless they chain into takeover). The CVE watchlist now spans 4,000+ product fingerprints (live from nuclei-templates when installed — refresh with sync_cve_corpus() before a big recon pass); live_recon's suggested_cves names the SPECIFIC CVE for the detected product/version — then cve_verify(cve, target) CONFIRMS it live by running the matching nuclei template (VERIFIED only when nuclei actually matches; a CVE is report-ready only after cve_verify, never from the fingerprint alone). For 'scan this target' automation, nuclei_scan(target) fingerprints the stack and runs the WHOLE matched template set (CVE templates + optional broad tech-tag) in one rate-limited pass — use it instead of verifying CVEs one-by-one. SURFACE-FIRST: surface_map(surface) maps an attack surface (image/redirect proxy, edge cache, payment portal, AI agent, file service, Electron, SSO) to its highest-value tests — map surfaces then test, don't spray payloads; then param_surface(target, class) discovers EVERY testable parameter (URL query, form fields, /users/123 path ids + UUIDs, JS refs) + per-class specialist hints — test the DISCOVERED params, not just the hinted one. PRECISION GUARD: resolve_check(host) detects when an ISP/country redirects the domain to a blocking/sinkhole page (gambling/adult/streaming niches) — audit the PUBLIC IP, never the sinkhole.\n" + "KNOWLEDGE BASE (discover, don't guess): skillLookup(topic) / listSkills for the 40 attack-path + hunting skills (ap-*, hp-*, bs-*); list_attack_vectors / attack_vectors(cat) for the 34-category vector taxonomy; taxonomy(kind) for OWASP/CWE/ASVS/API-top-10; read_playbook(type) for the 17 campaign playbooks; list_manuals / read_tool_manual for 317 tool references; list_tools / run_catalog_tool for the 141-tool catalog. Every tool, skill, and data layer is reachable from these. CROSS-TARGET INTELLIGENCE: strike_resolve auto-records confirmed/false_positive verdicts into the empirical ledger; query it with intelligence(action=hit_rates/top_vectors/recall) so a new target on a known tech inherits historical hit-rates (attack_plan already blends them into success_probability). intelligence(action=recall, tech, target) is the per-engagement pattern recall — it returns which patterns historically CONFIRMED on that tech (prioritize them) versus which only ever produced false positives (avoid them), so a new engagement starts from the empirical playbook of past ones. LIVE PENTEST + TARGETING (black-box, WAF-safe): cache_gap(base_url, endpoint) fuzzes path-normalization + static-extension variants for cache deception/poisoning; blind_oracle(base_url, param, value) detects which SINK a param reaches (sql/ssti/xss) via benign arithmetic (id=1+0, test${7*7}) WITHOUT a malicious payload; param_precedence(base_url, param, A, B) maps duplicate-parameter precedence (first/last-wins) for WAF-bypass; broken_link_scan(url) detects expired-domain takeover (crawls outbound refs → DNS/RDAP → flags claimable domains). Logic-bug live probers: jwt_forge(token, endpoint, …) forges JWT confusion tokens (alg:none / RS256→HS256 key-confusion / kid-jku-x5u) and proves which is ACCEPTED live; proto_pollution_probe(target, endpoint) sends __proto__ / constructor.prototype JSON canaries and reports VERIFIED only on reflection; smuggling_test(target) sends CL.TE / TE.CL desync probes over a raw connection. OOB (out-of-band) PROOF for BLIND vulns: oob_start() spins an interactsh listener (unique *.oast.* domain, streams DNS/HTTP interactions — the standard for public targets) or a local HTTP listener for local labs; inject the canary, then oob_poll(id) until the hit arrives; oob_stop(id) to clean up. A blind SSRF/XXE/SQLi/command-injection is CONFIRMED only when the canary actually arrives — never from a guess. SOURCE TARGETING: review_harness(path) FIRST for full coverage — it applies a fixed 5-signal rubric (entry_points/auth_gates/sinks/secrets/logic_bugs) to EVERY source file and sorts the signal-dense ones first (a complete map, not a partial scan); then action_surface(code) maps every wp_ajax_* action to its vuln class + DERIVED priority (what to look for WHERE); security_state(sanitizers, sink) gives a wrong-context sanitization verdict; framework_knowledge(framework) returns auth/nonce primitives for 9 frameworks; variant_scan(root, signature) mass-sweeps an install base for ONE bug family; patch_analyze(old, new) REVERSES a security patch into the 0-day, and its mine_signature feeds variant_scan to sweep unpatched copies. All of these are DETECTION HINTS — verify each live (strike_verify / browser_validate) before it enters the report. NEWEST DETECTORS — understand AND try them: variant_scan also emits priv_esc (privilege escalation — a role/capability/user change fed by request input, traced cross-file via reverse call-graph, e.g. set_role($id, sanitize_key($_POST['um-role'])) or update_user_meta(...,'wp_capabilities', $_POST[...])) and flags uploadAllow=>array('all') unrestricted-upload configs as file_upload (the file-manager/connector shape). TO TRY a priv_esc hit: submit the role field with 'administrator' (or send wp_capabilities meta) and confirm the account actually gains admin. TO TRY an unrestricted-upload hit: POST a .php/.phtml and confirm it lands web-reachable + executes. variant_scan also emits hardcoded_secret (CWE-798) in three tiers: a secret field (api_key/db_password/access_token/aws_*) assigned a literal (HIGH), a known secret FORMAT (JWT eyJ…/AWS AKIA…/GitHub ghp_…/Slack xoxb-…/…) regardless of name (HIGH), or a high-entropy suspect token (MEDIUM — verify before reporting). Verify a frontend/JS secret live via drive_devtools (inspect the JS bundle / network response and confirm the key is actually exposed), and for a backend secret confirm it is reachable/valid. A static hit is a HYPOTHESIS until verified live — never report unverified. For a repo, also scan_git_history(root) — secrets DELETED from HEAD still live in .git history (the removed-later-is-not-fixed trap), and it feeds every recovered pre-deletion blob through the same secret detector. Once you recover a credential, prove it is still ACTIVE with validate_leaked_key(apiBase, endpoints, key|blobUrl) — a READ-ONLY differential test (401-without-key vs 200-with-key = auth bypass PROVEN). Two supply-chain detectors: dependency_confusion (a manifest/build file resolving a package from a public registry — pypi.org/npmjs.org --extra-index-url, or a scoped @company/pkg with no private registry) and ml_supply_chain (torch.load without weights_only=True / joblib.load / np.load(allow_pickle=True) / keras load_model on an untrusted model → pickle RCE). Rust is supported end-to-end: the taint engine (blitz_scan / taint_file) maps Rust sinks (std::process::Command, sqlx/diesel, reqwest/hyper, std::fs, serde/bincode) to the classic classes, plus two Rust-specific detectors — rust_unsafe (transmute / assume_init / from_raw_parts / ptr::* — memory-unsafety surface, CWE-119) and rust_format_string (println!/format! with a variable as the format string — CWE-134). The sqlx query! compile-time macro is treated as safe (suppressed). Two more web-class detectors: jwt_alg_confusion (a JWT verifier that dispatches on the token's alg header and accepts BOTH RS256 and HS256, with the HS256 HMAC secret derived from public/shared key material — CWE-347; the attacker sets alg=HS256 and signs with the public key to forge admin tokens) and cache_deception (a reverse proxy caching static-extension URLs with a cache key that omits the session cookie, over a session-authenticated backend — CWE-525; the victim's cached private response is re-fetched anonymously). TO TRY a jwt_alg_confusion hit: fetch the public key (/.well-known/jwks.json), forge an HS256 token signed with that public key, and confirm admin access. TO TRY a cache_deception hit: request an authenticated path with a static extension appended (.css/.js) and re-fetch it anonymously. Three more detectors: cors_misconfiguration (Access-Control-Allow-Origin reflects the request Origin or is '*' while credentials are true — CWE-942; any site reads the victim's authenticated responses), xss (user input into a DOM sink — innerHTML / dangerouslySetInnerHTML / document.write / .html() / eval — CWE-79), and subdomain_takeover (a dangling CNAME/AliasTarget to a claimable service like *.github.io/*.herokuapp.com/*.amazonaws.com — CWE-404). TO TRY a cors hit: send an evil Origin + confirm the response echoes it with Allow-Credentials:true. TO TRY an xss hit: confirm the sink executes (drive_devtools). TO TRY a takeover hit: verify the CNAME target resolves to an unclaimed/not-found page. Two more client-side detectors: postmessage_no_origin_check (a window message listener that reads event.data but never validates event.origin — CWE-345; any site can drive the handler) and client_trusted_flag (a privilege/entitlement flag like isAdmin/withVip/premium read from request input and used in a security decision — CWE-807/639; derive the flag from server state, never the client). TO TRY a postMessage hit: send a crafted message from an attacker origin and confirm the handler acts on it. TO TRY a client-trusted-flag hit: flip the flag param to true and confirm the privilege is actually granted. WEB3: for Solidity/EVM, call web3_audit(code|path) — a deterministic 13-class smart-contract audit (reentrancy incl. ERC721/ERC1155 callback reentrancy, unchecked return, unchecked arithmetic, tx.origin auth, signature replay, unprotected selfdestruct, delegatecall-to-user-input, missing access control, timestamp dependence, spot-price oracle manipulation incl. slot0/latestRoundData, unbounded loop, abi.encodePacked hash collision, missing address(0) validation) returning ranked findings (class/CWE/severity/line), then foundry_poc(class, contract) to generate an executable Foundry test, and foundry_run(contract_code, poc_code|class_) to actually EXECUTE it — a PASSING PoC test is the proof that the exploit reproduces (deterministic: forge's own test verdict). A web3 hit is a HYPOTHESIS until the Foundry PoC reproduces it.\n" + `PRECISION ANALYSIS — chrome-devtools MCP is MANDATORY, not optional:
 ` + `- When a lead needs DOM/JS precision — DOM-XSS sink execution, AJAX endpoint + payload interception, JS runtime errors, redirect chains, SSO/token flow, or anything a static HTML glance cannot prove — call drive_devtools(target). It spawns chrome-devtools-mcp and drives a REAL headless browser (navigate -> list_network_requests -> list_console_messages -> evaluate_script), giving reproducible evidence instead of a guess.
 ` + `- Before calling it, check_mcp() to confirm the browser MCP is available; if not, blitzstrike install-tools or npm i -g chrome-devtools-mcp. Never substitute eyeballing minified JS or a static sink guess when a live browser can PROVE it. This is the precision + consistency layer.
 ` + "EXTERNAL MCP SERVERS (verified, connect them — NOT one-shot CLIs): chrome-devtools-mcp (29 tools: list_network_requests/get_network_request for raw HTTP interception, list_console_messages for JS errors + DOM-XSS, evaluate_script to run JS, take_screenshot) -> stdio chrome-devtools-mcp --headless --chromeArg=--no-sandbox --chromeArg=--disable-gpu (as root/container this is MANDATORY or the browser crashes; add --proxyServer http://127.0.0.1:8080 to route through Burp, --acceptInsecureCerts for bad TLS). burp-suite-mcp (proxy history + active scan + repeater) -> NOT npm: build the JAR (git clone https://github.com/PortSwigger/mcp-server.git && cd mcp-server && ./gradlew embedProxyJar), load it in Burp Extensions, then connect the client to http://127.0.0.1:9876 (SSE) or the stdio proxy java -jar mcp-proxy-all.jar --sse-url http://127.0.0.1:9876."
@@ -63973,7 +64072,7 @@ ${statusLine}` : statusLine },
     let src = code;
     if (!src && path) {
       try {
-        src = readFileSync31(path, "utf8");
+        src = readFileSync32(path, "utf8");
       } catch {
         return { content: [{ type: "text", text: JSON.stringify({ error: "unreadable path" }) }] };
       }
@@ -64054,7 +64153,7 @@ ${statusLine}` : statusLine },
     let src = code;
     if (!src && path) {
       try {
-        src = readFileSync31(path, "utf8");
+        src = readFileSync32(path, "utf8");
       } catch {
         return { content: [{ type: "text", text: JSON.stringify({ error: "unreadable path" }) }] };
       }
@@ -64189,6 +64288,16 @@ ${statusLine}` : statusLine },
     }
     return { content: [{ type: "text", text: JSON.stringify(r) }] };
   });
+  server.registerTool("param_surface", {
+    title: "Autonomous parameter discovery — extract every testable parameter",
+    description: "Fetch a target and extract EVERY testable parameter (URL query params, HTML form fields, path-segment identifiers /users/123 + UUIDs, JS-referenced endpoints) — then attach per-class discovery + test hints. Pass a class (xss/sqli/ssrf/idor/lfi/rce/xxe/jwt/open_redirect/prototype_pollution/csti/mass_assignment/header_injection/api_security/file_upload) to filter the specialist hints, or no class to get all 15. Test the discovered params (not just the hinted one) — e.g. an XSS hint about ?category= may hide a vulnerable searchTerm form field.",
+    inputSchema: {
+      target: string2().describe("Target URL, e.g. https://target.com"),
+      class_: string2().optional().describe("Vuln class to filter specialist hints (default: all)")
+    }
+  }, async ({ target, class_ }) => {
+    return { content: [{ type: "text", text: JSON.stringify(await paramSurface(target, class_)) }] };
+  });
   server.registerTool("jwt_forge", {
     title: "Forge JWT confusion tokens and verify which is accepted live",
     description: "Turn a captured JWT into forged attack tokens (alg:none, HS256 with empty key, HS256 with the public key = RS256→HS256 key confusion, plus kid/jku/x5u hints) and — when an endpoint is given — send each one to prove which is ACCEPTED (status < 400). A JWT hit is report-ready only when a forged token is accepted, never from header analysis alone.",
@@ -64855,7 +64964,7 @@ ${statusLine}` : statusLine },
     description: "Route confusion (separate class from source→sink taint): detect attacker-controlled dispatch (call_user_func/forward/dispatch), dynamic method calls ($obj->$m()), dynamic include (LFI), batch/proxy forwarding without per-route auth re-check, and route matching without path normalization.",
     inputSchema: { path: string2().describe("Source file path (PHP)") }
   }, async ({ path }) => {
-    const code = readFileSync31(path, "utf8");
+    const code = readFileSync32(path, "utf8");
     const findings = detectRouteConfusion(code, path);
     return {
       content: [{
@@ -64869,7 +64978,7 @@ ${statusLine}` : statusLine },
     description: "Complex-bug detection beyond taint: deserialization→POP gadget chain (unserialize + magic method), type juggling (loose ==/!= vs hash/secret), mass assignment (extract/parse_str without a safe flag), prototype pollution (unsafe merge of request data), CRLF/header injection (user input into a header), path confusion (user input into a file path without canonicalization), SSRF (user-controlled URL with no host allowlist), XXE (user XML parsed without disabling entities), SSTI (user input into a template render with no sandbox), wrong-context sanitization (sanitize_text_field() into a SQL query), missing authorization/nonce on WordPress entry points (wp_ajax_* / admin_post_* / register_rest_route without current_user_can, nonce verification, or permission_callback), CORS misconfiguration (reflected/wildcard origin + credentials), XSS (user input into a DOM sink: innerHTML / dangerouslySetInnerHTML / document.write / .html / eval), and subdomain takeover (CNAME/AliasTarget to a claimable service).",
     inputSchema: { path: string2().describe("Source file path") }
   }, async ({ path }) => {
-    const code = readFileSync31(path, "utf8");
+    const code = readFileSync32(path, "utf8");
     const findings = detectComplexBugs(code, path);
     return {
       content: [{
@@ -64955,7 +65064,7 @@ ${statusLine}` : statusLine },
     let lang = language;
     if (path && !code) {
       try {
-        src = readFileSync31(path, "utf8");
+        src = readFileSync32(path, "utf8");
       } catch {
         return { content: [{ type: "text", text: JSON.stringify({ error: `cannot read ${path}` }) }] };
       }
@@ -65204,15 +65313,15 @@ async function serve() {
 
 // src/index.ts
 init_cli();
-import { readFileSync as readFileSync35, existsSync as existsSync19 } from "node:fs";
-import { join as join33 } from "node:path";
-import { fileURLToPath as fileURLToPath21 } from "node:url";
-var ROOT14 = join33(fileURLToPath21(new URL(".", import.meta.url)), "..");
+import { readFileSync as readFileSync36, existsSync as existsSync19 } from "node:fs";
+import { join as join34 } from "node:path";
+import { fileURLToPath as fileURLToPath22 } from "node:url";
+var ROOT15 = join34(fileURLToPath22(new URL(".", import.meta.url)), "..");
 function readVersion() {
   try {
-    const p = join33(ROOT14, "package.json");
+    const p = join34(ROOT15, "package.json");
     if (existsSync19(p))
-      return JSON.parse(readFileSync35(p, "utf8")).version ?? "1.0.0";
+      return JSON.parse(readFileSync36(p, "utf8")).version ?? "1.0.0";
   } catch {}
   return "1.0.0";
 }

package/intelligence/specialist-discovery.json

@@ -0,0 +1,20 @@
+{
+  "_comment": "Specialist discovery heuristics: per vulnerability class, WHERE to discover the parameters to test and WHAT tool/technique to drive. Derived from public multi-agent scanner methodology (generalized — no source brand).",
+  "classes": [
+    { "id": "xss", "discover": "every reflected param (URL query + form field) + JS template sinks (dangerouslySetInnerHTML, innerHTML)", "test": "blind_oracle (sink detection) + drive_devtools (confirm execution)", "tools": ["blind_oracle", "drive_devtools", "browser_validate"] },
+    { "id": "sqli", "discover": "every query param + hidden form fields + CSRF tokens (SQLi can live in token validation)", "test": "blind_oracle (arithmetic id=1+0 vs id=1+1) + boolean/time inference", "tools": ["blind_oracle", "active_scan"] },
+    { "id": "ssrf", "discover": "URL-typed params (callback, url, redirect, webhook, img, proxy, fetch) + webhook-verification callbacks", "test": "blind_oracle + oob_start (canary URL must arrive)", "tools": ["blind_oracle", "oob_start", "oob_poll"] },
+    { "id": "idor", "discover": "path segments (/users/123) + UUIDs + sequential numeric ids in URLs", "test": "sibling_scan (differential IDOR attacker A → victim B)", "tools": ["sibling_scan", "active_scan"] },
+    { "id": "lfi", "discover": "file/read/path/include params (file, page, template, include, lang, path)", "test": "path traversal ../ sequences + encoding bypass", "tools": ["blind_oracle", "active_scan"] },
+    { "id": "rce", "discover": "command/code params (cmd, exec, ping, host, ip, eval) + deserialization endpoints", "test": "command injection i'''d / brace expansion + deserialization gadgets", "tools": ["blind_oracle", "oob_start"] },
+    { "id": "xxe", "discover": "XML upload endpoints + multipart forms + Content-Type: application/xml", "test": "XXE file disclosure + SSRF via XML parser + blind XXE via oob", "tools": ["oob_start", "oob_poll"] },
+    { "id": "jwt", "discover": "JWT patterns in URL params, body, cookies, localStorage, Authorization header (eyJ…)", "test": "jwt_analyze (parse + vectors) + jwt_forge (alg:none / RS256→HS256) live", "tools": ["jwt_analyze", "jwt_forge"] },
+    { "id": "open_redirect", "discover": "redirect/next/return/url params + OAuth redirect_uri", "test": "redirect to attacker-controlled URL (/%09/ // /\\ @ prefixes)", "tools": ["blind_oracle", "active_scan"] },
+    { "id": "prototype_pollution", "discover": "JSON POST acceptance (Content-Type: application/json) + merge/deep-extend endpoints", "test": "proto_pollution_probe (__proto__ / constructor.prototype canary)", "tools": ["proto_pollution_probe"] },
+    { "id": "csti", "discover": "Angular/Vue/React template sinks + {{ }} interpolation in reflected params", "test": "{{7*7}} / ${7*7} arithmetic reflection", "tools": ["blind_oracle", "drive_devtools"] },
+    { "id": "mass_assignment", "discover": "user-create/update endpoints + profile/payment POST bodies", "test": "inject role/is_admin/permissions/price/total fields", "tools": ["sibling_scan", "active_scan"] },
+    { "id": "header_injection", "discover": "header-reflecting params (host, x-forwarded-for, location, redirect)", "test": "CRLF injection + host-header cache poisoning", "tools": ["param_precedence", "active_scan"] },
+    { "id": "api_security", "discover": "REST/GraphQL routes + method tampering (GET↔POST↔PUT) + GraphQL introspection", "test": "horizontal/vertical privilege + method override", "tools": ["sibling_scan", "active_scan", "discover_api"] },
+    { "id": "file_upload", "discover": "multipart upload endpoints + avatar/import/attachment forms", "test": "extension/content-type bypass → web-reachable .php/.phtml", "tools": ["active_scan", "strike_verify"] }
+  ]
+}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "blitzstrike",
-  "version": "2.4.121",
+  "version": "2.4.122",
   "description": "Blitz Strike — a universal MCP security-audit toolbelt. BLITZ sweeps the attack surface, EAGLE-EYE traces source-to-sink, STRIKE verifies live. 57 attack chains, 130-tool catalog, intelligence data layer. One server, every agent.",
   "type": "module",
   "bin": {