@lincoln504/pi-research 1.3.10 → 1.3.12

package/dist/cli.mjs

@@ -269,7 +269,10 @@ var init_log_utils = __esm({
       `${B_LEFT}(sk-[A-Za-z0-9_-]{16,}|gh[posru]_[A-Za-z0-9]{16,}|github_pat_[A-Za-z0-9_]{20,}|AKIA[0-9A-Z]{16}|xox[baprs]-[A-Za-z0-9-]{10,})${B_RIGHT}`,
       "g"
     );
-    JWT_PATTERN = new RegExp(`${B_LEFT}eyJ[A-Za-z0-9_-]+\\.[A-Za-z0-9_-]+\\.[A-Za-z0-9_-]+`, "g");
+    JWT_PATTERN = new RegExp(
+      `${B_LEFT}eyJ[A-Za-z0-9_-]{1,1024}\\.[A-Za-z0-9_-]{1,8192}\\.[A-Za-z0-9_-]{1,4096}`,
+      "g"
+    );
     BASIC_AUTH_PATTERN = new RegExp(`${B_LEFT}Basic\\s+[A-Za-z0-9+/]{16,}={0,2}`, "gi");
     BEARER_AUTH_PATTERN = new RegExp(`${B_LEFT}Bearer\\s+[A-Za-z0-9._~+/-]+=*`, "gi");
     PROVIDER_TOKEN_PATTERN = new RegExp(
@@ -1158,8 +1161,16 @@ var init_service_registry = __esm({
        */
       async registerAndReplace(name, factory, options = {}) {
         const mergedOptions = { ...this.defaultOptions, ...options, allowOverwrite: true };
-        if (this.services.has(name)) {
-          const registration = this.services.get(name);
+        const registration = this.services.get(name);
+        if (registration) {
+          while (registration.disposalPromise) {
+            await registration.disposalPromise.catch(() => {
+            });
+          }
+          if (registration.initializationPromise) {
+            await registration.initializationPromise.catch(() => {
+            });
+          }
           if (registration.instance && registration.instance.dispose) {
             try {
               await registration.instance.dispose();
@@ -2148,13 +2159,18 @@ function isPlausibleCitationUrl(url) {
   }
 }
 function lastCitedLinksHeaderIndex(text) {
+  return lastCitedLinksHeader(text).textStart;
+}
+function lastCitedLinksHeader(text) {
   const re = /(^|\n)([ \t]*(?:#{1,6}[ \t]*|>[ \t]*|\*\*)?)CITED LINKS\b/gi;
-  let last = -1;
+  let lineStart = -1;
+  let textStart = -1;
   let m;
   while ((m = re.exec(text)) !== null) {
-    last = m.index + m[1].length + m[2].length;
+    lineStart = m.index + m[1].length;
+    textStart = m.index + m[1].length + m[2].length;
   }
-  return last;
+  return { lineStart, textStart };
 }
 function parseCitations(report) {
   const headerIdx = lastCitedLinksHeaderIndex(report);
@@ -2662,7 +2678,6 @@ function saveConfig(config, scope = "local", cwd = process.cwd(), changedKeys) {
     PI_RESEARCH_LLM_THINKING_LEVEL: config.LLM_THINKING_LEVEL,
     PI_RESEARCH_PLANNING_MAX_TOKENS: String(config.PLANNING_MAX_TOKENS),
     PI_RESEARCH_SYNTHESIS_MAX_TOKENS: String(config.SYNTHESIS_MAX_TOKENS),
-    PI_RESEARCH_RESEARCHER_MAX_TOKENS: String(config.RESEARCHER_MAX_TOKENS),
     PI_RESEARCH_MIGRATION_STRATEGY: config.MIGRATION_STRATEGY,
     PI_RESEARCH_CONSOLE_LOG: String(config.CONSOLE_LOG),
     ...config.RESEARCH_MODEL ? { PI_RESEARCH_MODEL: config.RESEARCH_MODEL } : {},
@@ -2827,7 +2842,6 @@ function createConfig(env, processEnv) {
     LLM_THINKING_LEVEL: parseEnvEnum(e, "PI_RESEARCH_LLM_THINKING_LEVEL", ["off", "minimal", "low", "medium", "high"], DEFAULTS.LLM_THINKING_LEVEL),
     PLANNING_MAX_TOKENS: parseEnvNumber(e, "PI_RESEARCH_PLANNING_MAX_TOKENS", DEFAULTS.PLANNING_MAX_TOKENS, 1024, 131072),
     SYNTHESIS_MAX_TOKENS: parseEnvNumber(e, "PI_RESEARCH_SYNTHESIS_MAX_TOKENS", DEFAULTS.SYNTHESIS_MAX_TOKENS, 1024, 131072),
-    RESEARCHER_MAX_TOKENS: parseEnvNumber(e, "PI_RESEARCH_RESEARCHER_MAX_TOKENS", DEFAULTS.RESEARCHER_MAX_TOKENS, 1024, 131072),
     MIGRATION_STRATEGY: parseEnvEnum(e, "PI_RESEARCH_MIGRATION_STRATEGY", ["drop", "re-embed", "backup"], DEFAULTS.MIGRATION_STRATEGY),
     CONSOLE_LOG: parseEnvBool(e, "PI_RESEARCH_CONSOLE_LOG", DEFAULTS.CONSOLE_LOG),
     RESEARCH_MODEL: parseEnvString(e, "PI_RESEARCH_MODEL", DEFAULTS.RESEARCH_MODEL),
@@ -2891,7 +2905,7 @@ function parseEnvNumber(env, key, def, min, max, integer = false) {
   return n;
 }
 function parseEnvBool(env, key, def) {
-  const v = env[key];
+  const v = env[key]?.trim();
   if (v === void 0 || v === "") return def;
   return v.toLowerCase() === "true";
 }
@@ -2899,7 +2913,7 @@ function parseEnvString(env, key, def) {
   return env[key] || def;
 }
 function parseEnvEnum(env, key, allowed, def) {
-  const raw = env[key];
+  const raw = env[key]?.trim();
   if (raw === void 0 || raw === "") return def;
   if (allowed.includes(raw)) return raw;
   logger.warn(`[config] Invalid ${key}="${raw}" (expected one of: ${allowed.join(", ")}). Using default "${def}".`);
@@ -3015,15 +3029,6 @@ var init_config = __esm({
        *  the model's real maxTokens at call time.
        *  Not exposed in TUI — controlled via PI_RESEARCH_SYNTHESIS_MAX_TOKENS env var. */
       SYNTHESIS_MAX_TOKENS: Type.Number({ minimum: 1024, maximum: 131072, default: 32768 }),
-      /** Max OUTPUT tokens per researcher LLM call (deep and quick sub-agent sessions).
-       *  Without a cap the session inherits the model's full catalog ceiling as the
-       *  per-request max_tokens — ~230k for large-context models — which providers that
-       *  budget-check max_tokens up front (OpenRouter 402s against remaining credits)
-       *  reject outright on low balances, even though the actual response would cost
-       *  cents. Default 16384: researcher turns are tool calls plus a cited report, far
-       *  below this. Clamped to the model's real ceiling at session build time.
-       *  Not exposed in TUI — controlled via PI_RESEARCH_RESEARCHER_MAX_TOKENS env var. */
-      RESEARCHER_MAX_TOKENS: Type.Number({ minimum: 1024, maximum: 131072, default: 16384 }),
       /** LLM Model override for researcher sub-agents and knowledge synthesis.
        *  Format: provider/model-id (e.g. google/gemini-2.0-flash-001) or just model-id.
        *  When set, this overrides ctx.model for researcher sub-agents (both deep and quick)
@@ -3060,7 +3065,8 @@ var init_config = __esm({
       /**
        * Comma-separated list of research tools to DISABLE for every researcher in a run
        * (e.g. "scrape", or "youtube_transcript,stackexchange"). Valid tool names:
-       * search, scrape, security_search, stackexchange, youtube_transcript, grep.
+       * search, scrape, security_search, stackexchange, youtube_transcript, grep, read
+       * (RESEARCH_TOOL_NAMES in constants.ts is the canonical list).
        * Disable-only: the list is merged into the per-run excludeTools stream, so disabled
        * tools are both removed from each researcher's toolset AND named in the
        * coordinator/evaluator "DISABLED TOOLS" prompt section. Env-only
@@ -3101,7 +3107,6 @@ var init_config = __esm({
       "PI_RESEARCH_LLM_THINKING_LEVEL",
       "PI_RESEARCH_PLANNING_MAX_TOKENS",
       "PI_RESEARCH_SYNTHESIS_MAX_TOKENS",
-      "PI_RESEARCH_RESEARCHER_MAX_TOKENS",
       "PI_RESEARCH_MIGRATION_STRATEGY",
       "PI_RESEARCH_CONSOLE_LOG",
       "PI_RESEARCH_MODEL",
@@ -3146,7 +3151,7 @@ function getMaxScrapeBatches(config) {
 function resolveExcludedTools(callerExclusions, configDisabled = []) {
   return [.../* @__PURE__ */ new Set([...DEFAULT_EXCLUDED_TOOLS, ...callerExclusions ?? [], ...configDisabled])];
 }
-var MAX_GATHERING_CALLS, MAX_GREP_CALLS, MAX_SCRAPE_URLS, MAX_TEAM_SIZE_LEVEL_1, MAX_TEAM_SIZE_LEVEL_2, MAX_TEAM_SIZE_LEVEL_3, MAX_ROUNDS_LEVEL_1, MAX_ROUNDS_LEVEL_2, MAX_ROUNDS_LEVEL_3, MAX_EXTRA_ROUNDS_WITH_STEERING, OSV_TIMEOUT_MS, WEBGPU_PROBE_TIMEOUT_MS, MAX_QUERY_LENGTH, MAX_FILENAME_QUERY_LENGTH, MIN_QUERY_LENGTH, MAX_EXPORT_RETRIES, DEFAULT_MAX_RETRIES, DEFAULT_INITIAL_DELAY_MS, DEFAULT_MAX_DELAY_MS, DEFAULT_MODEL_CONTEXT_WINDOW, BATCH_2_DEFAULT_CONCURRENCY, MAX_SCRAPE_CONTENT_CHARS_PER_DOC, RESEARCHER_LAUNCH_DELAY_MS, MAX_QUERIES_PER_RESEARCHER_LEVEL_1, MAX_QUERIES_PER_RESEARCHER_LEVEL_2, MAX_QUERIES_PER_RESEARCHER_LEVEL_3, DEFAULT_EXCLUDED_TOOLS;
+var MAX_GATHERING_CALLS, MAX_GREP_CALLS, MAX_SCRAPE_URLS, MAX_TEAM_SIZE_LEVEL_1, MAX_TEAM_SIZE_LEVEL_2, MAX_TEAM_SIZE_LEVEL_3, MAX_ROUNDS_LEVEL_1, MAX_ROUNDS_LEVEL_2, MAX_ROUNDS_LEVEL_3, MAX_EXTRA_ROUNDS_WITH_STEERING, OSV_TIMEOUT_MS, WEBGPU_PROBE_TIMEOUT_MS, MAX_QUERY_LENGTH, MAX_FILENAME_QUERY_LENGTH, MIN_QUERY_LENGTH, MAX_EXPORT_RETRIES, DEFAULT_MAX_RETRIES, DEFAULT_INITIAL_DELAY_MS, DEFAULT_MAX_DELAY_MS, DEFAULT_MODEL_CONTEXT_WINDOW, BATCH_2_DEFAULT_CONCURRENCY, MAX_SCRAPE_CONTENT_CHARS_PER_DOC, RESEARCHER_LAUNCH_DELAY_MS, MAX_QUERIES_PER_RESEARCHER_LEVEL_1, MAX_QUERIES_PER_RESEARCHER_LEVEL_2, MAX_QUERIES_PER_RESEARCHER_LEVEL_3, DEFAULT_EXCLUDED_TOOLS, RESEARCH_TOOL_NAMES;
 var init_constants = __esm({
   "src/constants.ts"() {
     "use strict";
@@ -3178,6 +3183,15 @@ var init_constants = __esm({
     MAX_QUERIES_PER_RESEARCHER_LEVEL_2 = 15;
     MAX_QUERIES_PER_RESEARCHER_LEVEL_3 = 20;
     DEFAULT_EXCLUDED_TOOLS = ["grep"];
+    RESEARCH_TOOL_NAMES = [
+      "search",
+      "scrape",
+      "security_search",
+      "stackexchange",
+      "youtube_transcript",
+      "grep",
+      "read"
+    ];
   }
 });
 
@@ -5395,6 +5409,9 @@ var init_store = __esm({
               logger.warn("[store] Failed to read table schema:", schemaErr);
             }
           } else {
+            if (this.options.embedder.getDimension() === null) {
+              await this.withEmbedderReconnect((embedder) => embedder.embedMany([" "]));
+            }
             this.table = await this.createTable();
           }
           await this.evictOldRecords();
@@ -5776,6 +5793,45 @@ var init_store = __esm({
           }
         }
       }
+      /**
+       * Refresh the timestamp of one CURRENT generation for url+type — the
+       * revalidation "touch" for the writer-queue's content-unchanged skip. Without
+       * it a stable page kept its original timestamp forever: with
+       * KNOWLEDGE_STORE_MAX_SERVE_AGE_DAYS set, the serve-age gate then treated the
+       * entry as permanently stale (a live re-scrape on every run, for exactly the
+       * content that never changes), and TTL eviction eventually deleted an entry
+       * that had been re-verified byte-identical possibly hours earlier.
+       */
+      async touchByUrlAndType(url, ingestionType, generationTs, newTs) {
+        return this.trackOperation("touchByUrlAndType", void 0, async () => {
+          let retryCount = 0;
+          const MAX_RETRIES = 5;
+          const BASE_DELAY = 100;
+          const scopeFilter = this.getScopeFilter();
+          while (retryCount <= MAX_RETRIES) {
+            const table = await this.getFreshTable();
+            try {
+              const escapedUrl = url.replace(/'/g, "''");
+              const escapedType = ingestionType.replace(/'/g, "''");
+              await table.update({
+                where: `url = '${escapedUrl}' AND ingestion_type = '${escapedType}' AND timestamp = ${Math.floor(generationTs)} AND (${scopeFilter})`,
+                valuesSql: { timestamp: String(Math.floor(newTs)) }
+              });
+              logger.log(`[store] Refreshed timestamp for unchanged ${ingestionType} chunks of ${url}`);
+              return;
+            } catch (err) {
+              if (isLanceCommitConflict(err)) {
+                retryCount++;
+                if (retryCount > MAX_RETRIES) throw err;
+                const delay2 = Math.floor(Math.random() * (BASE_DELAY * Math.pow(2, retryCount - 1)));
+                await new Promise((r) => setTimeout(r, delay2));
+                continue;
+              }
+              throw err;
+            }
+          }
+        });
+      }
       async findDocumentsByUrl(url) {
         return this.trackOperation("findDocumentsByUrl", [], async () => {
           const table = await this.getFreshTable();
@@ -6013,7 +6069,10 @@ var init_store = __esm({
               backupDirs.map(async (name) => ({ name, ts: await this.backupDirTimestamp(name) }))
             );
             ranked.sort((a, b) => a.ts - b.ts);
-            for (const { name } of ranked.slice(0, -1)) {
+            const ttlDays = this.options.ttlDays ?? 30;
+            const backupCutoff = Math.trunc(Date.now() - ttlDays * 24 * 60 * 60 * 1e3);
+            for (const { name, ts } of ranked.slice(0, -1)) {
+              if (ts >= backupCutoff) continue;
               const fullPath = path13.join(this.options.dbDir, name);
               logger.info(`[store] Removing superseded backup directory: ${name}`);
               await fsPromises.rm(fullPath, { recursive: true, force: true, maxRetries: 5, retryDelay: 50 });
@@ -6309,6 +6368,17 @@ var init_writer_queue = __esm({
                 metrics.increment("knowledge_ingest_stale_prune_failed_total", 1);
               });
             }
+            if (this.options.store.touchByUrlAndType) {
+              try {
+                const recheck = (await this.options.store.findByUrl(normalizedUrl)).filter((c) => c.metadata["ingestionType"] === incomingType);
+                const stillNewest = recheck.length > 0 && Math.max(...recheck.map((d) => Number(d.timestamp) || 0)) === newestTs;
+                if (stillNewest) {
+                  await this.options.store.touchByUrlAndType(normalizedUrl, incomingType, newestTs, Date.now());
+                }
+              } catch (err) {
+                logger.warn(`[writer-queue] Timestamp refresh for unchanged ${normalizedUrl} (${incomingType}) failed:`, err);
+              }
+            }
             logger.log(`[writer-queue] Skipping ${normalizedUrl} (${incomingType}) \u2014 content unchanged.`);
             return;
           }
@@ -6787,7 +6857,17 @@ async function loadPipelineWithTimeout(model, device, timeoutMs, useCache) {
         interOpNumThreads: 1
       }
     });
-    return await withTimeout3(pipelinePromise, timeoutMs, errorMessage);
+    let timedOut = true;
+    try {
+      const p = await withTimeout3(pipelinePromise, timeoutMs, errorMessage);
+      timedOut = false;
+      return p;
+    } finally {
+      if (timedOut) {
+        pipelinePromise.then((late) => late.dispose?.()).catch(() => {
+        });
+      }
+    }
   });
   return { pipeline: loadedPipeline, errorMessage };
 }
@@ -6916,7 +6996,17 @@ async function loadModelOnCPU(model, initializationTimeoutMs, useCache) {
         interOpNumThreads: 1
       }
     });
-    return await withTimeout3(pipelinePromise, initializationTimeoutMs, "CPU fallback model load timed out");
+    let timedOut = true;
+    try {
+      const p = await withTimeout3(pipelinePromise, initializationTimeoutMs, "CPU fallback model load timed out");
+      timedOut = false;
+      return p;
+    } finally {
+      if (timedOut) {
+        pipelinePromise.then((late) => late.dispose?.()).catch(() => {
+        });
+      }
+    }
   });
   return loadedPipeline;
 }
@@ -9113,6 +9203,7 @@ function formatQueriesBlock(queries) {
 }
 function showKnowledgeSearchWidget(ctx) {
   if (ctx.mode !== "tui" || !ctx.hasUI) return;
+  knowledgeWidgetRefCount++;
   try {
     const panelFactory = createKnowledgeSearchPanel();
     ctx.ui.setWidget(KNOWLEDGE_WIDGET_ID, panelFactory, { placement: "aboveEditor" });
@@ -9122,6 +9213,8 @@ function showKnowledgeSearchWidget(ctx) {
 }
 function hideKnowledgeSearchWidget(ctx) {
   if (ctx.mode !== "tui" || !ctx.hasUI) return;
+  knowledgeWidgetRefCount = Math.max(0, knowledgeWidgetRefCount - 1);
+  if (knowledgeWidgetRefCount > 0) return;
   try {
     ctx.ui.setWidget(KNOWLEDGE_WIDGET_ID, void 0, { placement: "aboveEditor" });
   } catch (err) {
@@ -9601,7 +9694,7 @@ function createResearchKnowledgeSearchTool(iface) {
     }
   };
 }
-var ResearchKnowledgeSearchParams, MAX_REFERENCE_CHARS, MAX_DOCUMENTS, MAX_CANDIDATES, TRIAGE_DESCRIPTION_MAX_CHARS, TRIAGE_MAX_TOKENS, KNOWLEDGE_WIDGET_ID, RESEARCH_KNOWLEDGE_MISS_STRING, RESEARCH_KNOWLEDGE_MAYBE_STRING, KNOWLEDGE_SYNTHESIS_MAX_ATTEMPTS, KNOWLEDGE_SYNTHESIS_RETRY_BASE_MS;
+var ResearchKnowledgeSearchParams, MAX_REFERENCE_CHARS, MAX_DOCUMENTS, MAX_CANDIDATES, TRIAGE_DESCRIPTION_MAX_CHARS, TRIAGE_MAX_TOKENS, KNOWLEDGE_WIDGET_ID, knowledgeWidgetRefCount, RESEARCH_KNOWLEDGE_MISS_STRING, RESEARCH_KNOWLEDGE_MAYBE_STRING, KNOWLEDGE_SYNTHESIS_MAX_ATTEMPTS, KNOWLEDGE_SYNTHESIS_RETRY_BASE_MS;
 var init_research_knowledge_search = __esm({
   "src/tools/research-knowledge-search.ts"() {
     "use strict";
@@ -9636,6 +9729,7 @@ var init_research_knowledge_search = __esm({
     TRIAGE_DESCRIPTION_MAX_CHARS = 600;
     TRIAGE_MAX_TOKENS = 2048;
     KNOWLEDGE_WIDGET_ID = "pi-research-knowledge-search";
+    knowledgeWidgetRefCount = 0;
     RESEARCH_KNOWLEDGE_MISS_STRING = "No results found. Live research can get the info.";
     RESEARCH_KNOWLEDGE_MAYBE_STRING = "Partial results found in knowledge store. Live research can fill gaps.";
     KNOWLEDGE_SYNTHESIS_MAX_ATTEMPTS = 3;
@@ -10432,7 +10526,7 @@ function normalizeCitations(reports) {
         localToGlobal.set(localKey, globalId);
       }
     });
-    const headerIdx = lastCitedLinksHeaderIndex(report);
+    const headerIdx = lastCitedLinksHeader(report).lineStart;
     let content = headerIdx >= 0 ? report.slice(0, headerIdx) : report;
     content = rewriteCitationMarkers(
       content,
@@ -11391,9 +11485,13 @@ import { platform, homedir as homedir2 } from "node:os";
 function getWindowsCamoufoxDir() {
   return join7(homedir2(), "AppData", "Local", "camoufox", "camoufox", "Cache");
 }
+function getCustomCamoufoxDir() {
+  return process.env["CAMOUFOX_INSTALL_DIR"] || process.env["PLAYWRIGHT_BROWSERS_PATH"] || void 0;
+}
 function getBrowserCacheDir() {
-  if (process.env["PLAYWRIGHT_BROWSERS_PATH"]) {
-    return process.env["PLAYWRIGHT_BROWSERS_PATH"];
+  const custom = getCustomCamoufoxDir();
+  if (custom) {
+    return custom;
   }
   const osPlatform = platform();
   if (osPlatform === "win32") {
@@ -11424,11 +11522,13 @@ function getBrowserProfileDir(config) {
 }
 function getBrowserEnv(config) {
   const env = { ...process.env };
-  const customPath = process.env["PLAYWRIGHT_BROWSERS_PATH"];
+  const customPath = getCustomCamoufoxDir();
   if (customPath) {
     env["PLAYWRIGHT_BROWSERS_PATH"] = customPath;
+    env["CAMOUFOX_INSTALL_DIR"] = customPath;
   } else {
     delete env["PLAYWRIGHT_BROWSERS_PATH"];
+    delete env["CAMOUFOX_INSTALL_DIR"];
   }
   if (!env["PI_RESEARCH_LOG_FILE"]) {
     const logFilePath = getLogger().getLogFilePath();
@@ -11457,7 +11557,7 @@ function ensureBrowserCacheDir() {
   return cacheDir;
 }
 function getCamoufoxBinaryPath() {
-  const customPath = process.env["PLAYWRIGHT_BROWSERS_PATH"];
+  const customPath = getCustomCamoufoxDir();
   if (customPath) {
     return customPath;
   }
@@ -14235,8 +14335,22 @@ var GPUResourceService = class {
   name = ServiceNames.GPU_RESOURCE_SERVICE;
   lifecycle = "uninitialized" /* UNINITIALIZED */;
   processLifecycle = null;
-  gpuLockStaleThresholdMs = 1e3 * 60 * 5;
-  // 5 minutes default
+  /**
+   * Backstop for a holder that is alive but wedged. Step 3 already reclaims from
+   * *dead* owners via PID+startTime liveness, so this only ever fires against a
+   * process we just proved is running — which makes it a lock STEAL, and the
+   * threshold must therefore exceed the longest legitimate hold.
+   *
+   * It previously did not: at 5 minutes it exactly equalled the default
+   * EMBEDDING_MODEL_INIT_TIMEOUT_MS (300000, schema max 600000), and the
+   * embedder holds this lock across the whole model download + pipeline load.
+   * A first-time HuggingFace fetch running near its own permitted budget would
+   * have the GPU lock taken out from under it, admitting exactly the concurrent
+   * WebGPU/Dawn init this lock exists to prevent. Keep this strictly above that
+   * schema maximum.
+   */
+  gpuLockStaleThresholdMs = 1e3 * 60 * 15;
+  // 15 min > EMBEDDING_MODEL_INIT_TIMEOUT_MS max (10 min)
   constructor(options) {
     if (options) {
       this.processLifecycle = options.processLifecycle;
@@ -17046,19 +17160,33 @@ function isSsrfBlockError(error) {
   return false;
 }
 var cachedSsrfFetcher = void 0;
-async function getSsrfSafeFetcher() {
-  if (cachedSsrfFetcher !== void 0) return cachedSsrfFetcher;
+function getSsrfSafeFetcher() {
+  cachedSsrfFetcher ??= (async () => {
+    try {
+      const { Agent: Agent2, fetch: undiciFetch } = await import("undici");
+      return {
+        fetch: undiciFetch,
+        dispatcher: new Agent2({ connect: { lookup: ssrfSafeLookup } })
+      };
+    } catch (e) {
+      logger.warn(`[Scrapers] undici unavailable; connect-time SSRF pinning disabled (request-time validation still active): ${e instanceof Error ? e.message : String(e)}`);
+      return null;
+    }
+  })();
+  return cachedSsrfFetcher;
+}
+async function disposeSsrfSafeFetcher() {
+  const pending = cachedSsrfFetcher;
+  cachedSsrfFetcher = void 0;
+  if (pending === void 0) return;
+  const cached = await pending.catch(() => null);
+  const closable = cached?.dispatcher;
+  if (!closable?.close) return;
   try {
-    const { Agent: Agent2, fetch: undiciFetch } = await import("undici");
-    cachedSsrfFetcher = {
-      fetch: undiciFetch,
-      dispatcher: new Agent2({ connect: { lookup: ssrfSafeLookup } })
-    };
+    await closable.close();
   } catch (e) {
-    logger.warn(`[Scrapers] undici unavailable; connect-time SSRF pinning disabled (request-time validation still active): ${e instanceof Error ? e.message : String(e)}`);
-    cachedSsrfFetcher = null;
+    logger.debug(`[Scrapers] Error closing SSRF dispatcher: ${e instanceof Error ? e.message : String(e)}`);
   }
-  return cachedSsrfFetcher;
 }
 function formatErrorWithCause(err) {
   try {
@@ -17917,7 +18045,6 @@ import { Value as Value7 } from "typebox/value";
 // src/security/nvd-client.ts
 init_retry_utils();
 init_circuit_breaker();
-init_safe_unref();
 init_metrics();
 init_http_body();
 
@@ -17980,7 +18107,6 @@ var NVDRateLimiter = class {
           signal?.removeEventListener("abort", onAbort);
           resolve9();
         }, waitTime);
-        safeUnref(timeoutId);
         const onAbort = () => {
           clearTimeout(timeoutId);
           reject(Object.assign(new Error("NVD search aborted"), { name: "AbortError" }));
@@ -18202,6 +18328,7 @@ async function fetchPaginated(term, options, maxResults, severityVersion, admitE
   const signal = options?.signal;
   let startIndex = 0;
   let totalPagesFetched = 0;
+  let seekedToNewest = false;
   while (totalPagesFetched < maxPages && allVulnerabilities.length < maxResults) {
     if (signal?.aborted) break;
     const url = buildURL(term, options, pageSize, startIndex, severityVersion);
@@ -18222,9 +18349,17 @@ async function fetchPaginated(term, options, maxResults, severityVersion, admitE
       break;
     }
     metrics.increment("nvd_cache_hits_total", 1);
-    allVulnerabilities.push(...admitted);
     const responseData = data;
-    if (responseData?.["totalResults"] !== void 0 && typeof responseData["totalResults"] === "number" && startIndex + pageSize >= responseData["totalResults"]) {
+    const reportedTotal = typeof responseData?.["totalResults"] === "number" ? responseData["totalResults"] : void 0;
+    if (!seekedToNewest && startIndex === 0 && reportedTotal !== void 0 && reportedTotal > maxResults) {
+      seekedToNewest = true;
+      startIndex = reportedTotal - maxResults;
+      allVulnerabilities.length = 0;
+      totalPagesFetched++;
+      continue;
+    }
+    allVulnerabilities.push(...admitted);
+    if (reportedTotal !== void 0 && startIndex + pageSize >= reportedTotal) {
       break;
     }
     startIndex += pageSize;
@@ -18232,6 +18367,14 @@ async function fetchPaginated(term, options, maxResults, severityVersion, admitE
   }
   metrics.increment("nvd_pagination_requests_total", totalPagesFetched);
   metrics.observe("nvd_pagination_pages_fetched", totalPagesFetched);
+  allVulnerabilities.sort((a, b) => {
+    const ta = a.published ? Date.parse(a.published) : NaN;
+    const tb = b.published ? Date.parse(b.published) : NaN;
+    if (Number.isNaN(ta) && Number.isNaN(tb)) return 0;
+    if (Number.isNaN(ta)) return 1;
+    if (Number.isNaN(tb)) return -1;
+    return tb - ta;
+  });
   return allVulnerabilities.slice(0, maxResults);
 }
 function shouldSupplementWithV2(options) {
@@ -19242,7 +19385,14 @@ async function searchOSV(terms, options) {
         uniqueVulns.set(vuln.id, vuln);
       }
     }
-    vulnerabilities.push(...Array.from(uniqueVulns.values()));
+    vulnerabilities.push(...Array.from(uniqueVulns.values()).sort((a, b) => {
+      const ta = a.published ? Date.parse(a.published) : NaN;
+      const tb = b.published ? Date.parse(b.published) : NaN;
+      if (Number.isNaN(ta) && Number.isNaN(tb)) return 0;
+      if (Number.isNaN(ta)) return 1;
+      if (Number.isNaN(tb)) return -1;
+      return tb - ta;
+    }));
     if (failures.length > 0 && failures.length === terms.length) {
       throw new Error(`All OSV lookups failed: ${failures.join("; ")}`);
     }
@@ -21549,8 +21699,6 @@ async function createResearcherSession(options) {
     cwd
   });
   logger.info(`[Researcher] Using model for researcher ${researcherId}: ${modelToUse.provider}/${modelToUse.id}`);
-  const outputCap = config?.RESEARCHER_MAX_TOKENS ?? 16384;
-  const sessionModel = modelToUse.maxTokens > outputCap ? { ...modelToUse, maxTokens: outputCap } : modelToUse;
   try {
     const tracker = new ToolUsageTracker(createDefaultToolLimits(config));
     const globalLinks = updateGlobalLinks || (() => {
@@ -21591,7 +21739,7 @@ async function createResearcherSession(options) {
       sessionManager: SessionManager.inMemory(),
       // Each researcher gets its own isolated session
       settingsManager: researcherSettings,
-      model: sessionModel,
+      model: modelToUse,
       // pi 0.80.8 replaced the SDK's `modelRegistry`/`authStorage` session options
       // with the async `modelRuntime`. Pass the runtime backing OUR registry when
       // this run built one (standalone CLI/SDK): it carries any explicit API key
@@ -21622,7 +21770,7 @@ async function createResearcherSession(options) {
         throw new Error("Session creation returned invalid result");
       }
       sessionRef.session = result.session;
-      return { session: result.session, resolvedModel: sessionModel };
+      return { session: result.session, resolvedModel: modelToUse };
     } catch (tailError) {
       const session = result?.session;
       if (session && typeof session.abort === "function") {
@@ -23091,6 +23239,17 @@ var ResearchOrchestrationService = class {
     const { ctx, query, depth = 0, observer, onUpdate, sessionId, researchId, config, excludeTools } = options;
     const researchConfig = config || getConfig(ctx.cwd);
     const configDisabledTools = researchConfig.DISABLED_TOOLS ? researchConfig.DISABLED_TOOLS.split(",").map((s) => s.trim()).filter(Boolean) : [];
+    for (const [surface, names] of [
+      ["excludeTools", excludeTools ?? []],
+      ["PI_RESEARCH_DISABLED_TOOLS", configDisabledTools]
+    ]) {
+      const unknown = names.filter((t) => !RESEARCH_TOOL_NAMES.includes(t));
+      if (unknown.length > 0) {
+        logger.warn(
+          `[ResearchOrchestrationService] ${surface}: ignoring unknown tool name${unknown.length === 1 ? "" : "s"} ${unknown.map((t) => `"${t}"`).join(", ")} \u2014 these exclude nothing. Valid names: ${RESEARCH_TOOL_NAMES.join(", ")}.`
+        );
+      }
+    }
     const mergedExcludeTools = resolveExcludedTools(excludeTools, configDisabledTools);
     const selectedModel = await this.resolveResearchModel(options);
     logger.info(`[ResearchOrchestrationService] Using model: ${selectedModel.provider}/${selectedModel.id}`);
@@ -23104,6 +23263,13 @@ var ResearchOrchestrationService = class {
       });
       logger.info(`[ResearchOrchestrationService] Acquired research run slot ${runSlot.slotIndex} (cap ${semaphore.getMaxSlots()}).`);
     } catch (err) {
+      const isTerminal = err instanceof ResearchRunCapacityError || err instanceof Error && err.name === "AbortError";
+      if (isTerminal) {
+        try {
+          observer?.onError?.(err);
+        } catch {
+        }
+      }
       if (err instanceof ResearchRunCapacityError) throw err;
       if (err instanceof Error && err.name === "AbortError") throw err;
       logger.warn(`[ResearchOrchestrationService] Run-cap unavailable, proceeding without it: ${err instanceof Error ? err.message : String(err)}`);
@@ -23855,7 +24021,7 @@ Research performed using ${modelId}`;
           if (globalId !== void 0) localToGlobal.set(localKey, globalId);
           else droppedNumbers.add(localKey);
         });
-        let body = synthesis.slice(0, citedHeaderIdx).trimEnd();
+        let body = synthesis.slice(0, lastCitedLinksHeader(synthesis).lineStart).trimEnd();
         if (localToGlobal.size > 0 || droppedNumbers.size > 0) {
           body = rewriteCitationMarkers(body, localToGlobal, (n) => droppedNumbers.has(n));
         }
@@ -24502,6 +24668,12 @@ async function _doShutdown() {
     logger.error("[SDK] Error clearing session state:", err);
     errors.push(err instanceof Error ? err : new Error(String(err)));
   }
+  try {
+    await disposeSsrfSafeFetcher();
+  } catch (err) {
+    logger.error("[SDK] Error disposing SSRF fetcher:", err);
+    errors.push(err instanceof Error ? err : new Error(String(err)));
+  }
   if (globalContainer) {
     try {
       await disposeCoreServices(globalContainer);
@@ -24778,6 +24950,7 @@ Total cost: ${costStr}`;
 // src/cli.ts
 init_config();
 init_model_registry_factory();
+init_constants();
 import { getAgentDir as getAgentDir2 } from "@earendil-works/pi-coding-agent";
 function readPkgVersion() {
   try {
@@ -24898,14 +25071,15 @@ function makeProgressObserver() {
 var ConfigFileError = class extends Error {
 };
 function bridgeConfigEnv(explicitConfigPath) {
-  const basePath = explicitConfigPath ?? getGlobalEnvFilePath();
+  const globalBasePath = getGlobalEnvFilePath();
+  const basePath = explicitConfigPath ?? globalBasePath;
   const cliOverlayPath = getInterfaceEnvFilePath("cli");
   let loaded = false;
   const merged = {};
   const fromExplicitBase = /* @__PURE__ */ new Set();
-  const layers = explicitConfigPath !== void 0 ? [cliOverlayPath, basePath] : [basePath, cliOverlayPath];
+  const layers = explicitConfigPath !== void 0 ? [globalBasePath, cliOverlayPath, explicitConfigPath] : [globalBasePath, cliOverlayPath];
   for (const filePath of layers) {
-    const isExplicit = explicitConfigPath !== void 0 && filePath === basePath;
+    const isExplicit = explicitConfigPath !== void 0 && filePath === explicitConfigPath;
     try {
       if (!fs17.existsSync(filePath)) {
         if (isExplicit) throw new ConfigFileError(`--config file not found: ${filePath}`);
@@ -25042,7 +25216,11 @@ async function detectCredentials(explicitModel) {
     modelFrom,
     piAuthPresent,
     piModelsPresent,
-    problem: "No model or API key is configured for pi-research. Provide credentials via environment variables (PI_RESEARCH_API_KEY + PI_RESEARCH_PROVIDER + PI_RESEARCH_MODEL) or, if you use pi, via pi's configuration (a provider entry in ~/.pi/agent/auth.json, or a per-provider apiKey in ~/.pi/agent/models.json)."
+    // A model configured with no usable key is a different (and misleading to
+    // conflate) gap than nothing configured at all: SKILL.md tells agents to relay
+    // this text verbatim, so "No model … is configured" pointed users who HAD set
+    // PI_RESEARCH_MODEL at the wrong knob.
+    problem: model ? `A model is configured (${model}) but no API key is available for ${provider ? `provider "${provider}"` : "its provider"}. Provide one via the PI_RESEARCH_API_KEY environment variable (with PI_RESEARCH_PROVIDER), or, if you use pi, via pi's configuration (a provider entry in ~/.pi/agent/auth.json, or a per-provider apiKey in ~/.pi/agent/models.json).` : "No model or API key is configured for pi-research. Provide credentials via environment variables (PI_RESEARCH_API_KEY + PI_RESEARCH_PROVIDER + PI_RESEARCH_MODEL) or, if you use pi, via pi's configuration (a provider entry in ~/.pi/agent/auth.json, or a per-provider apiKey in ~/.pi/agent/models.json)."
   };
 }
 function configBlock(det, extraNote) {
@@ -25112,6 +25290,8 @@ Error: ${vmsg}
       ...args.excludeTools ? { excludeTools: args.excludeTools } : {},
       ...args.initialLinks ? { initialLinks: args.initialLinks } : {}
     }, abortController.signal);
+    const outcome = getLastResearcherOutcome();
+    const errorReport = getLastErrorReport();
     let reportPath = null;
     const exportCfg = getConfig(process.cwd(), "cli");
     if (exportCfg.RESEARCH_REPORT_EXPORT_ENABLED) {
@@ -25126,8 +25306,6 @@ Error: ${vmsg}
       exit = cancellationSignal ? exitCodeForSignal(cancellationSignal) : EXIT.CANCELLED;
     }
     if (args.json) {
-      const outcome = getLastResearcherOutcome();
-      const errorReport = getLastErrorReport();
       toStdout(pretty({
         ok: true,
         depth,
@@ -25204,8 +25382,11 @@ Configure in:
     if (activeResearchAbortController === abortController) activeResearchAbortController = null;
     await safeShutdown();
   }
+  if (cancellationRequested) {
+    exit = cancellationSignal ? exitCodeForSignal(cancellationSignal) : EXIT.CANCELLED;
+  }
   if (json) {
-    toStdout(pretty({ ok: true, ...result }));
+    toStdout(pretty({ ok: true, ...cancellationRequested ? { cancelled: true } : {}, ...result }));
   } else {
     toStdout(result.text.endsWith("\n") ? result.text : result.text + "\n");
   }
@@ -25215,6 +25396,7 @@ async function cmdStatus(json) {
   const det = await detectCredentials();
   const paths = resolvedConfigPaths();
   const cfg = getConfig(process.cwd(), "cli");
+  const browserInstalled = isBrowserAvailable();
   const summary = {
     package: "@lincoln504/pi-research",
     version: PKG_VERSION,
@@ -25228,6 +25410,12 @@ async function cmdStatus(json) {
       modelFrom: det.modelFrom,
       problem: det.problem ?? null
     },
+    browser: {
+      installed: browserInstalled,
+      // Same remedy the health check prints, so both surfaces name one command.
+      fix: browserInstalled ? null : "npx camoufox-js fetch",
+      note: browserInstalled ? null : "Stealth browser not installed \u2014 scraping falls back to plain fetch, which some sites block. npm 11.19+ blocks install scripts by default, which skips this package's postinstall."
+    },
     knowledgeStoreMode: cfg.KNOWLEDGE_STORE_MODE,
     defaultDepth: cfg.DEFAULT_RESEARCH_DEPTH,
     paths
@@ -25242,6 +25430,13 @@ async function cmdStatus(json) {
 
 `);
   toStdout(configBlock(det, summary.ready ? void 0 : "\u2014 fix the problem above, then re-run.") + "\n");
+  if (!summary.browser.installed) {
+    toStdout(`
+stealth browser: NOT INSTALLED \u2014 run: ${summary.browser.fix}
+`);
+    toStdout(`  ${summary.browser.note}
+`);
+  }
   toStdout(`
 knowledge store mode: ${cfg.KNOWLEDGE_STORE_MODE}   (default depth: ${cfg.DEFAULT_RESEARCH_DEPTH})
 `);
@@ -25440,7 +25635,7 @@ async function reportError(err, what, json) {
   const isRateLimit = !isCancellation && (lower.includes("429") || lower.includes("rate limit") || lower.includes("too many requests"));
   const exitCode = isCancellation ? cancellationSignal ? exitCodeForSignal(cancellationSignal) : EXIT.CANCELLED : isCapacity ? EXIT.TEMPFAIL : isConfigError && !isRateLimit ? EXIT.CONFIG : EXIT.SOFTWARE;
   if (json) {
-    const payload = { ok: false, error: msg2, exitCode };
+    const payload = isCancellation ? { ok: false, error: `pi-research ${what} cancelled.`, cause: msg2, exitCode } : { ok: false, error: msg2, exitCode };
     if ((isCapacity || isRateLimit) && !isCancellation) payload["retryable"] = true;
     if (isCancellation) payload["cancelled"] = true;
     if (process.env["PI_RESEARCH_DEBUG"] === "true" && err instanceof Error && err.stack) {
@@ -25540,6 +25735,7 @@ COMMANDS
     --dry-run | --json           Plan only; JSON output.
 
   help, --help, -h               Show this help.
+  --version, -v                  Print the pi-research version.
 
 CONFIGURE
   Credentials are resolved in this order (first match wins):
@@ -25729,6 +25925,12 @@ function parseArgs(argv) {
       } else if (a === "--exclude-tools") {
         const v = takeOptionValue(rest, ++i, "--exclude-tools");
         excludeTools = v.split(",").map((s) => s.trim()).filter(Boolean);
+        const unknown = excludeTools.filter((t) => !RESEARCH_TOOL_NAMES.includes(t));
+        if (unknown.length > 0) {
+          throw new UsageError(
+            `--exclude-tools: unknown tool name${unknown.length === 1 ? "" : "s"} ${unknown.map((t) => `"${t}"`).join(", ")}. Valid names: ${RESEARCH_TOOL_NAMES.join(", ")}.`
+          );
+        }
         if (excludeTools.length === 0) excludeTools = void 0;
       } else if (a === "--initial-links") {
         let j = i + 1;
@@ -25742,6 +25944,7 @@ function parseArgs(argv) {
           }
           j++;
         }
+        if (j === i + 1) throw new UsageError("--initial-links requires at least one URL.");
         i = j - 1;
       } else if (a === "--config") {
         configPath = takeOptionValue(rest, ++i, "--config");