@jtalk22/slack-mcp 4.4.1 → 4.4.3

package/README.md

@@ -60,7 +60,7 @@ Six prebuilt templates ship with the package:
 npx -y @jtalk22/slack-mcp --apply-template oncall-handoff --channels C012345,C067890
 ```
 
-Available templates: `oncall-handoff`, `support-triage`, `exec-monday`, `sprint-tracker`, `customer-feedback`, `incident-room`. The structural primitives (`slack_workflow_save`, `slack_workflows`) are free forever in OSS; the hosted brain is `$0` to start (no card) and `$9/mo` Pro for unlimited AI tools (scheduled morning catch-up DM in development).
+Available templates: `oncall-handoff`, `support-triage`, `exec-monday`, `sprint-tracker`, `customer-feedback`, `incident-room`. The structural primitives (`slack_workflow_save`, `slack_workflows`) are free forever in OSS; the hosted brain is `$0` to start (no card) and `$19/mo` Pro for unlimited AI tools.
 
 ## Quick Start per Client
 
@@ -268,10 +268,10 @@ Hosted tiers at [mcp.revasserlabs.com](https://mcp.revasserlabs.com):
 | Tier | Price | What it owns |
 |------|-------|-------------|
 | Self-host | Free (MIT) | Local stdio, all 21 tools (16 read/write Slack + 2 workflow profile primitives + 3 discoverable upgrade stubs to hosted brain) |
-| Hosted Free | $0 (no card) | Email signup, 1 workspace, 10 smart_search/mo + 3 catch_me_up/mo + 5 triage/day. All 5 workflow profile types. 7-day index retention. |
-| Pro | $9/mo | Unlimited AI tools, **scheduled morning catch-up DM** *(in development, 8am workspace tz)*, permanent OAuth, 90-day Vectorize, 2 workspaces |
-| Team | $49/mo flat | Pro + shared workflow profiles + audit log + 24h support + scheduled catch-up to channel + 5 workspaces |
-| Ops | from $199/mo (custom) | SLA, custom retention, SOC2 evidence path, multi-tenant isolation, 10+ workspaces, dedicated workflow tuning |
+| Hosted Free | $0 (no card) | Email signup, 1 workspace, 2,000 requests/mo + 25 AI tool calls/mo. All 5 workflow profile types. 7-day index retention. |
+| Pro | $19/mo or $190/yr | Unlimited requests (fair use), unlimited AI tool calls, permanent OAuth, email support, 2 workspaces |
+| Team | $49/mo or $490/yr flat | Everything in Pro + shared workflow profiles, 5 workspaces, 24h support |
+| Safeguard | $199/mo — waitlist | Agent approval gates, scheduled catch-up DM, workspace memory — all *(in development)*. Waitlist only. |
 
 </details>
 
@@ -324,6 +324,22 @@ Full release notes on [GitHub releases/latest](https://github.com/jtalk22/slack-
 
 </details>
 
+## Token expired? / OAuth Lifeboat
+
+Session tokens (`xoxc-` + `xoxd-`) are extracted from your browser, and Slack rotates them roughly every 1-2 weeks. When they die, every tool call fails to authenticate. Instead of surfacing a raw Slack error, this server detects token death — `invalid_auth`, `not_authed`, `token_expired`, `token_revoked`, `account_inactive`, or an HTTP 401 — and returns a recovery message at the moment of pain.
+
+**Self-fix — re-extract fresh tokens:**
+
+```bash
+npx -y @jtalk22/slack-mcp --setup
+```
+
+On macOS with a logged-in Slack tab open in Chrome, you can instead call the `slack_refresh_tokens` tool (or run `npm run tokens:auto`). To avoid silent expiration during long idle windows, set up the optional [token-refresh LaunchAgent](docs/SETUP.md#keep-tokens-fresh-while-claude-is-closed-macos-optional).
+
+**Permanent fix — no rotation:** the [hosted version](https://mcp.revasserlabs.com/setup?utm_source=lifeboat&utm_medium=npm&utm_campaign=token_death) uses OAuth, which does not rotate every 1-2 weeks. Free tier available, no card.
+
+The recovery message appears at most once per process per hour; repeat failures inside that window get a one-line reminder so agents in retry loops don't spam. Set `SLACK_MCP_NO_UPSELL=1` to drop the hosted-option line while keeping the self-fix guidance.
+
 ## Rich Message Fields
 
 Added in 4.4.0. The four read tools marked ‡ above accept `include_rich_message_fields: true`, which surfaces the parts of a message that live outside `text` — `attachments`, `blocks`, `files`, `reactions`, `metadata`, plus `subtype`/`bot_id`/`app_id` (automated/bot/app markers) and `team` (workspace id).
@@ -410,4 +426,4 @@ Not affiliated with Slack Technologies, Inc. Uses browser session credentials
 
 ---
 
-Hosted version live at [mcp.revasserlabs.com](https://mcp.revasserlabs.com): Free tier (no card), $9/mo Pro, $49/mo Team flat, Ops from $199/mo. Hosted owns the AI brain (smart_search, catch_me_up, triage), the scheduled morning catch-up DM at 8am workspace time *(in development)*, permanent OAuth (no 2-week token rotation), 90-day Vectorize retention, and shared workflow profiles. The OSS package owns local stdio + the 16 Slack tools (12 read, 4 write) + workflow profile primitives (slack_workflow_save, slack_workflows). The 3 paid stubs (slack_smart_search, slack_catch_me_up, slack_triage) appear in OSS as discoverable upgrade prompts.
+Hosted version live at [mcp.revasserlabs.com](https://mcp.revasserlabs.com): Free tier (no card — 2,000 requests/mo + 25 AI tool calls/mo), $19/mo Pro (unlimited, permanent OAuth), $49/mo Team flat, and Safeguard $199/mo (waitlist). Hosted owns the AI brain (smart_search, catch_me_up, triage), permanent OAuth (no 2-week token rotation), and shared workflow profiles; the scheduled morning catch-up DM at 8am workspace time is a Safeguard feature *(in development)*. The OSS package owns local stdio + the 16 Slack tools (12 read, 4 write) + workflow profile primitives (slack_workflow_save, slack_workflows). The 3 paid stubs (slack_smart_search, slack_catch_me_up, slack_triage) appear in OSS as discoverable upgrade prompts.

package/docs/DEPLOYMENT-MODES.md

@@ -8,7 +8,7 @@ Use this guide to choose the right operating mode before rollout.
 - Choose local `web` for browser workflows and manual Slack browsing.
 - Choose hosted HTTP only when you need remote execution and can handle token operations.
 - Choose Smithery/Worker only when your consumers require registry-hosted MCP transport.
-- A managed **Hosted** version is live at [mcp.revasserlabs.com](https://mcp.revasserlabs.com) — Free tier (no card), Pro $9/mo, Team $49/mo flat, Ops from $199/mo. See [pricing](https://mcp.revasserlabs.com/pricing).
+- A managed **Hosted** version is live at [mcp.revasserlabs.com](https://mcp.revasserlabs.com) — Free tier (no card), Pro $19/mo, Team $49/mo flat, Safeguard $199/mo (waitlist). See [pricing](https://mcp.revasserlabs.com/pricing).
 
 ## Mode Matrix
 

package/docs/TROUBLESHOOTING.md

@@ -31,7 +31,7 @@ If `--version` fails here, the issue is install/runtime path, not Slack credenti
 
 ## Hosted Version
 
-The hosted version is live at [mcp.revasserlabs.com](https://mcp.revasserlabs.com). Free tier (no card) ships 10 smart_search/mo + 3 catch_me_up/mo + 5 triage/day + all 5 workflow profile types. Pro at $9/mo unlocks unlimited AI tools, the scheduled morning catch-up DM at 8am workspace time *(in development)*, permanent OAuth (no 2-week token rotation), and 90-day Vectorize retention. Team at $49/mo flat covers 5 workspaces with shared workflow profiles and audit log. Ops engagement starts at $199/mo (custom) for 10+ workspace organizations with SLA, custom retention, SOC2 evidence, or multi-tenant isolation.
+The hosted version is live at [mcp.revasserlabs.com](https://mcp.revasserlabs.com). Free tier (no card) ships 2,000 requests/mo + 25 AI tool calls/mo + all 5 workflow profile types. Pro at $19/mo (or $190/yr) unlocks unlimited requests and AI tool calls, permanent OAuth (no 2-week token rotation), email support, and 2 workspaces. Team at $49/mo flat (or $490/yr) covers 5 workspaces with shared workflow profiles and 24h support. Safeguard at $199/mo (waitlist only) adds agent approval gates, the scheduled morning catch-up DM at 8am workspace time, and workspace memory — all *(in development)*.
 
 The OSS package keeps the local-machine path. The hosted version adds the AI brain (smart_search, catch_me_up, triage) — these tools also appear in the OSS package as discoverable upgrade stubs that point at the hosted signup.
 

package/lib/handlers.js

@@ -21,6 +21,7 @@ import {
   ALLOWED_WORKFLOW_KINDS_LIST,
 } from "./workflow-store.js";
 import { withRichMessageFields } from "./rich-message-fields.js";
+import { isAuthDeath, buildLifeboatPayload } from "./lifeboat.js";
 
 // ============ Utilities ============
 
@@ -169,6 +170,11 @@ export async function handleHealthCheck() {
       token_updated: creds.updatedAt || null
     });
   } catch (e) {
+    // OAuth Lifeboat: a dead session token here is the most common first
+    // signal of token death — hand back recovery guidance, not a bare error.
+    if (isAuthDeath(e)) {
+      return asMcpJson(buildLifeboatPayload(e), true);
+    }
     return asMcpJson({
       status: "error",
       code: "auth_failed",
@@ -336,14 +342,17 @@ export async function handleListConversations(args) {
 export async function handleConversationsHistory(args) {
   const resolveUsers = args.resolve_users !== false;
   const includeRichMessageFields = parseBool(args.include_rich_message_fields);
-  const result = await slackAPI("conversations.history", {
+  const historyParams = {
     channel: args.channel_id,
     limit: args.limit || 50,
     oldest: args.oldest,
     latest: args.latest,
-    inclusive: true,
     include_all_metadata: parseBool(args.include_all_metadata)
-  });
+  };
+  // Only opt into boundary-inclusive reads when a boundary was actually
+  // provided — otherwise leave Slack's default behavior untouched.
+  if (args.oldest || args.latest) historyParams.inclusive = true;
+  const result = await slackAPI("conversations.history", historyParams);
 
   const messages = await Promise.all((result.messages || []).map(async (msg) => {
     const userName = resolveUsers ? await resolveUser(msg.user) : msg.user;
@@ -384,15 +393,18 @@ export async function handleGetFullConversation(args) {
 
   // Fetch all messages with pagination
   while (hasMore && allMessages.length < maxMessages) {
-    const result = await slackAPI("conversations.history", {
+    const historyParams = {
       channel: args.channel_id,
       limit: Math.min(100, maxMessages - allMessages.length),
       oldest: args.oldest,
       latest: args.latest,
       cursor,
-      inclusive: true,
       include_all_metadata: parseBool(args.include_all_metadata)
-    });
+    };
+    // Boundary-inclusive only when a boundary was actually provided —
+    // otherwise leave Slack's default behavior untouched.
+    if (args.oldest || args.latest) historyParams.inclusive = true;
+    const result = await slackAPI("conversations.history", historyParams);
 
     for (const msg of result.messages || []) {
       const userName = await resolveUser(msg.user);
@@ -738,13 +750,32 @@ export async function handleConversationsUnreads(args) {
 /**
  * Search users handler - client-side filter on users.list
  */
+// Explicit scan cap for client-side user search: stop paginating after
+// scanning this many workspace users (or when the cursor is exhausted) and
+// flag the result as truncated so total_matches is never misreported as
+// complete.
+const USERS_SEARCH_MAX_SCANNED = 1000;
+
 export async function handleUsersSearch(args) {
-  const query = (args.query || "").toLowerCase();
+  const rawQuery = typeof args.query === "string" ? args.query : "";
+  const query = rawQuery.trim().toLowerCase();
   const limit = args.limit || 20;
 
-  // Fetch all users (paginated)
+  // An empty/whitespace query would match every user in the workspace.
+  if (!query) {
+    return asMcpJson({
+      status: "error",
+      code: "invalid_arguments",
+      message: "query must be a non-empty string (an empty query would match all users).",
+      next_action: "Provide a name, display name, real name, or email fragment to search for."
+    }, true);
+  }
+
+  // Fetch users (paginated) and filter client-side
   const allUsers = [];
   let cursor;
+  let scannedUsers = 0;
+  let truncated = false;
 
   do {
     const result = await slackAPI("users.list", {
@@ -753,6 +784,7 @@ export async function handleUsersSearch(args) {
     });
 
     for (const u of (result.members || [])) {
+      scannedUsers++;
       if (u.deleted || u.is_bot || u.id === "USLACKBOT") continue;
 
       const searchFields = [
@@ -776,13 +808,18 @@ export async function handleUsersSearch(args) {
     }
 
     cursor = result.response_metadata?.next_cursor;
+    if (cursor && scannedUsers >= USERS_SEARCH_MAX_SCANNED) {
+      truncated = true;
+      break;
+    }
     if (cursor) await sleep(100);
-  } while (cursor && allUsers.length < 500);
+  } while (cursor);
 
   return asMcpJson({
     query: args.query,
     count: Math.min(allUsers.length, limit),
     total_matches: allUsers.length,
+    truncated,
     users: allUsers.slice(0, limit)
   });
 }
@@ -840,8 +877,8 @@ const HOSTED_UPGRADE_PAYLOAD = {
   message: "This tool needs hosted mode (Vectorize + Workers AI). Get free monthly credits at mcp.revasserlabs.com — no card required.",
   signup_url: "https://mcp.revasserlabs.com/signup",
   upgrade_url: "https://mcp.revasserlabs.com/pricing",
-  free_tier_quota: "10 smart_search + 3 catch_me_up per month, 5 triage per day",
-  pro_value_prop: "Pro $9/mo unlocks unlimited AI tools (scheduled morning catch-up DM at 8am workspace time rolling out Q2 2026).",
+  free_tier_quota: "2,000 requests/mo + 25 AI tool calls/mo (no card)",
+  pro_value_prop: "Pro $19/mo unlocks unlimited requests and AI tool calls, permanent OAuth (no token rotation).",
 };
 
 export async function handleSmartSearch(args) {
@@ -873,3 +910,35 @@ export async function handleTriage(args) {
     true
   );
 }
+
+// ============ Shared Tool Dispatch Map ============
+// Single source of truth mapping every advertised tool name (lib/tools.js)
+// to its handler. Both transports — stdio (src/server.js) and HTTP
+// (src/server-http.js) — dispatch through this map so the advertised tool
+// list and the dispatch surface cannot drift apart.
+
+export const TOOL_HANDLERS = Object.freeze({
+  slack_token_status: handleTokenStatus,
+  slack_health_check: handleHealthCheck,
+  slack_refresh_tokens: handleRefreshTokens,
+  slack_list_conversations: handleListConversations,
+  slack_conversations_history: handleConversationsHistory,
+  slack_get_full_conversation: handleGetFullConversation,
+  slack_search_messages: handleSearchMessages,
+  slack_users_info: handleUsersInfo,
+  slack_send_message: handleSendMessage,
+  slack_get_thread: handleGetThread,
+  slack_list_users: handleListUsers,
+  slack_add_reaction: handleAddReaction,
+  slack_remove_reaction: handleRemoveReaction,
+  slack_conversations_mark: handleConversationsMark,
+  slack_conversations_unreads: handleConversationsUnreads,
+  slack_users_search: handleUsersSearch,
+  // Workflow profile primitives (OSS local JSON store)
+  slack_workflow_save: handleWorkflowSave,
+  slack_workflows: handleWorkflows,
+  // Hosted-only AI tools (OSS = upgrade stubs)
+  slack_smart_search: handleSmartSearch,
+  slack_catch_me_up: handleCatchMeUp,
+  slack_triage: handleTriage,
+});

package/lib/lifeboat.js

@@ -0,0 +1,171 @@
+/**
+ * OAuth Lifeboat
+ *
+ * Detects Slack session-token death at the moment of pain and returns a
+ * genuinely helpful recovery message instead of a raw Slack API error.
+ *
+ * Session tokens (xoxc/xoxd) are extracted from the browser and Slack rotates
+ * them roughly every 1-2 weeks. When they die, every tool call fails with an
+ * auth error. This module classifies that failure and hands the user the
+ * self-fix first (re-extract) and the permanent fix second (hosted OAuth).
+ *
+ * Design rules:
+ * - Self-fix is always shown; the hosted option is shown second and is
+ *   suppressed entirely by SLACK_MCP_NO_UPSELL=1.
+ * - The long-form message is emitted at most once per process per hour;
+ *   subsequent failures inside that window get a one-line version so agents
+ *   in retry loops do not spam.
+ * - Never reads or logs token values — it only inspects error codes.
+ */
+
+// Slack API error codes (and the HTTP 401 case) that mean the session token
+// is dead and cannot be auto-healed — the point where the Lifeboat fires.
+export const AUTH_DEATH_SLACK_CODES = new Set([
+  "invalid_auth",
+  "not_authed",
+  "token_expired",
+  "token_revoked",
+  "account_inactive",
+]);
+
+const THROTTLE_WINDOW_MS = 60 * 60 * 1000; // one long-form message per process per hour
+
+const SETUP_CMD = "npx -y @jtalk22/slack-mcp --setup";
+const README_ANCHOR_URL =
+  "https://github.com/jtalk22/slack-mcp-server#token-expired--oauth-lifeboat";
+const HOSTED_SETUP_URL =
+  "https://mcp.revasserlabs.com/setup?utm_source=lifeboat&utm_medium=npm&utm_campaign=token_death";
+
+// Module-level throttle state (shared across every transport in one process).
+let lastLongFormAt = 0;
+
+/**
+ * Reset the throttle. Intended for tests.
+ */
+export function resetLifeboatThrottle() {
+  lastLongFormAt = 0;
+}
+
+function upsellEnabled() {
+  return process.env.SLACK_MCP_NO_UPSELL !== "1";
+}
+
+/**
+ * Pull the specific Slack auth-death code out of an error, if present.
+ * Returns the code string (e.g. "token_revoked") or null.
+ */
+export function authDeathCode(err) {
+  if (!err) return null;
+
+  if (typeof err === "string") {
+    const s = err.trim().toLowerCase();
+    return AUTH_DEATH_SLACK_CODES.has(s) ? s : null;
+  }
+
+  const slackCode =
+    typeof err.slack_error === "string" ? err.slack_error.trim().toLowerCase() : null;
+  if (slackCode && AUTH_DEATH_SLACK_CODES.has(slackCode)) return slackCode;
+
+  const msg = typeof err.message === "string" ? err.message.trim().toLowerCase() : "";
+  if (AUTH_DEATH_SLACK_CODES.has(msg)) return msg;
+  for (const code of AUTH_DEATH_SLACK_CODES) {
+    if (msg.includes(code)) return code;
+  }
+
+  return null;
+}
+
+/**
+ * Classify an error as AUTH-DEATH.
+ * Accepts a thrown Error (optionally carrying `.slack_error`, `.code`,
+ * `.status`), a raw Slack error-code string, or an HTTP-401 signal.
+ */
+export function isAuthDeath(err) {
+  if (!err) return false;
+
+  // The wrapped auth-failure lib/slack-client.js throws after a failed auto-heal.
+  if (typeof err === "object" && err.code === "token_auth_failed") return true;
+
+  if (authDeathCode(err)) return true;
+
+  // HTTP 401 (defensive — Slack normally returns HTTP 200 with ok:false).
+  if (typeof err === "string") {
+    const s = err.trim().toLowerCase();
+    return s === "401" || s.includes("http 401") || s.includes("status 401");
+  }
+  const status = err.status ?? err.statusCode ?? err.httpStatus ?? null;
+  if (status === 401) return true;
+  const msg = typeof err.message === "string" ? err.message.toLowerCase() : "";
+  return msg.includes("http 401") || msg.includes("status 401");
+}
+
+/**
+ * Build the structured recovery payload for an auth-death error.
+ * First call within the throttle window returns the long form; later calls
+ * return the one-line form. Honors SLACK_MCP_NO_UPSELL.
+ */
+export function buildLifeboatPayload(err, options = {}) {
+  const now = options.now ?? Date.now();
+  const slackError = authDeathCode(err);
+  const upsell = upsellEnabled();
+
+  const throttled = lastLongFormAt > 0 && now - lastLongFormAt < THROTTLE_WINDOW_MS;
+
+  if (throttled) {
+    const message = upsell
+      ? `Slack session token is still expired or revoked. Re-extract with \`${SETUP_CMD}\`. Permanent fix (hosted OAuth, free tier): ${HOSTED_SETUP_URL}`
+      : `Slack session token is still expired or revoked. Re-extract with \`${SETUP_CMD}\`.`;
+    return {
+      status: "error",
+      code: "slack_auth_expired",
+      slack_error: slackError,
+      throttled: true,
+      message,
+      next_action: `Run \`${SETUP_CMD}\`.`,
+    };
+  }
+
+  lastLongFormAt = now;
+
+  const payload = {
+    status: "error",
+    code: "slack_auth_expired",
+    slack_error: slackError,
+    throttled: false,
+    message:
+      "Your Slack session token has expired or been revoked. Slack rotates browser session tokens (xoxc/xoxd) roughly every 1-2 weeks, so this is expected — the server can no longer authenticate until you supply fresh tokens.",
+    self_fix:
+      `Re-extract fresh tokens: run \`${SETUP_CMD}\`. On macOS with Slack open in Chrome you can instead call the slack_refresh_tokens tool. Full steps: ${README_ANCHOR_URL}`,
+  };
+
+  // Preserve the auto-heal diagnostic when the wrapped error carried one.
+  if (err && typeof err === "object" && err.extraction_error) {
+    payload.extraction_error = err.extraction_error;
+  }
+
+  if (upsell) {
+    payload.hosted_option =
+      `To stop re-extracting every 1-2 weeks, switch to hosted OAuth (it never rotates): ${HOSTED_SETUP_URL} — free tier available, no credit card.`;
+    payload.next_action = `Run \`${SETUP_CMD}\` to re-extract tokens, or switch to permanent hosted OAuth (see hosted_option).`;
+  } else {
+    payload.next_action = `Run \`${SETUP_CMD}\` to re-extract tokens (macOS: call the slack_refresh_tokens tool).`;
+  }
+
+  return payload;
+}
+
+/**
+ * Wrap the recovery payload as an MCP tool-error response.
+ * Used by both transports (stdio + HTTP) so the recovery surface is identical.
+ */
+export function lifeboatResponse(err) {
+  return {
+    content: [
+      {
+        type: "text",
+        text: JSON.stringify(buildLifeboatPayload(err), null, 2),
+      },
+    ],
+    isError: true,
+  };
+}

package/lib/public-pages.js

@@ -63,7 +63,7 @@ function shareLinks() {
 }
 
 function shareNote() {
-  return `<strong>Verify in 30 seconds:</strong> <code>--version</code>, <code>--doctor</code>, <code>--status</code>. Self-host gives ${PUBLIC_METADATA.selfHostedToolCount} tools with session-based auth. Works with any MCP client — Claude, ChatGPT, Cursor, Copilot, Gemini, Windsurf. Hosted free tier (no card) live at <a href="${PUBLIC_METADATA.canonicalSiteUrl}">mcp.revasserlabs.com</a> — Pro $9/mo unlocks unlimited AI tools (scheduled morning catch-up DM rolling out Q2 2026).`;
+  return `<strong>Verify in 30 seconds:</strong> <code>--version</code>, <code>--doctor</code>, <code>--status</code>. Self-host gives ${PUBLIC_METADATA.selfHostedToolCount} tools with session-based auth. Works with any MCP client — Claude, ChatGPT, Cursor, Copilot, Gemini, Windsurf. Hosted free tier (no card) live at <a href="${PUBLIC_METADATA.canonicalSiteUrl}">mcp.revasserlabs.com</a> — Pro $19/mo unlocks unlimited AI tools and permanent OAuth.`;
 }
 
 function demoLinks() {
@@ -75,7 +75,7 @@ function demoLinks() {
 }
 
 function demoNote() {
-  return `Self-host free for ${PUBLIC_METADATA.selfHostedToolCount} tools with session-based auth. Works with Claude, ChatGPT, Cursor, Copilot, Gemini, Windsurf, and any other MCP client. No OAuth app, no admin approval. Hosted free tier (no card) live at <a href="${PUBLIC_METADATA.canonicalSiteUrl}" target="_blank" rel="noopener noreferrer">mcp.revasserlabs.com</a> — Pro $9/mo unlocks unlimited AI tools (scheduled morning catch-up DM rolling out Q2 2026).`;
+  return `Self-host free for ${PUBLIC_METADATA.selfHostedToolCount} tools with session-based auth. Works with Claude, ChatGPT, Cursor, Copilot, Gemini, Windsurf, and any other MCP client. No OAuth app, no admin approval. Hosted free tier (no card) live at <a href="${PUBLIC_METADATA.canonicalSiteUrl}" target="_blank" rel="noopener noreferrer">mcp.revasserlabs.com</a> — Pro $19/mo unlocks unlimited AI tools and permanent OAuth.`;
 }
 
 function demoFooterLinks() {
@@ -113,7 +113,7 @@ function commonTokens() {
     SELF_HOSTED_TOOL_COUNT: String(PUBLIC_METADATA.selfHostedToolCount),
     CLOUD_MANAGED_TOOL_COUNT: "15",
     TEAM_AI_WORKFLOW_COUNT: "3",
-    CLOUD_SOLO_PRICE: "$9/mo",
+    CLOUD_SOLO_PRICE: "$19/mo",
     CLOUD_TEAM_PRICE: "$49/mo",
     CLOUD_TURNKEY_LAUNCH_PRICE: "contact us",
     CLOUD_MANAGED_RELIABILITY_PRICE: "contact us",

package/lib/token-store.js

@@ -8,7 +8,7 @@
  * 4. Chrome auto-extraction (fallback)
  */
 
-import { readFileSync, writeFileSync, existsSync, renameSync, unlinkSync, chmodSync, copyFileSync, mkdtempSync, statSync, readdirSync } from "fs";
+import { readFileSync, writeFileSync, existsSync, renameSync, unlinkSync, rmSync, chmodSync, copyFileSync, mkdtempSync, statSync, readdirSync } from "fs";
 import { homedir, platform, tmpdir } from "os";
 import { join } from "path";
 import { execFileSync } from "child_process";
@@ -265,13 +265,22 @@ function extractCookieForProfile(profileDir) {
   const tmpDb = join(tmpDir, 'Cookies');
   try {
     copyFileSync(cookiesPath, tmpDb);
+    // Chrome keeps recent writes in the SQLite WAL sidecar until the next
+    // checkpoint — copy Cookies-wal (and -shm) too when present so the
+    // snapshot isn't stale.
+    for (const suffix of ['-wal', '-shm']) {
+      const sidecarPath = `${cookiesPath}${suffix}`;
+      if (existsSync(sidecarPath)) {
+        try { copyFileSync(sidecarPath, `${tmpDb}${suffix}`); } catch {}
+      }
+    }
 
     const queryResult = execFileSync('sqlite3', [
       tmpDb,
       "SELECT hex(encrypted_value) FROM cookies WHERE host_key LIKE '%.slack.com%' AND name = 'd' LIMIT 1;"
     ], { encoding: 'utf-8', timeout: 5000 }).trim();
 
-    try { unlinkSync(tmpDb); unlinkSync(tmpDir); } catch {}
+    try { rmSync(tmpDir, { recursive: true, force: true }); } catch {}
 
     if (!queryResult) return null;
 
@@ -304,7 +313,7 @@ function extractCookieForProfile(profileDir) {
     if (xoxdIndex < 0) return null;
     return text.substring(xoxdIndex);
   } catch {
-    try { unlinkSync(tmpDb); unlinkSync(tmpDir); } catch {}
+    try { rmSync(tmpDir, { recursive: true, force: true }); } catch {}
     return null;
   }
 }
@@ -358,7 +367,10 @@ function extractTokenFromLevelDB(profileDir) {
       const txt = readFileSync(f.path).toString('binary');
       XOXC_TOKEN_RE.lastIndex = 0;
       const matches = txt.match(XOXC_TOKEN_RE);
-      if (matches && matches.length) return matches[0];
+      // LevelDB .log/.ldb files append newer records after older ones, so
+      // the LAST match in a file is the most recently cached token — the
+      // first match can be a stale, already-rotated token.
+      if (matches && matches.length) return matches[matches.length - 1];
     } catch {
       continue;
     }

package/lib/tools.js

@@ -504,7 +504,7 @@ export const TOOLS = [
   // pointing at mcp.revasserlabs.com — the hosted worker actually runs them.
   {
     name: "slack_smart_search",
-    description: "Semantic + lexical hybrid search across your indexed Slack history. Returns ranked results with relevance scores, channel context, thread context, and matched terms. Hosted-only (requires Vectorize + Workers AI). Free tier ships 10 calls/month; upgrade to Pro $9/mo for unlimited at mcp.revasserlabs.com/pricing.",
+    description: "Semantic + lexical hybrid search across your indexed Slack history. Returns ranked results with relevance scores, channel context, thread context, and matched terms. Hosted-only (requires Vectorize + Workers AI). Free tier ships 25 AI tool calls/month (shared across the hosted AI tools); upgrade to Pro $19/mo for unlimited at mcp.revasserlabs.com/pricing.",
     inputSchema: {
       type: "object",
       properties: {
@@ -537,7 +537,7 @@ export const TOOLS = [
   },
   {
     name: "slack_catch_me_up",
-    description: "Run a structured catch-up against a saved workflow profile. Returns structured JSON per the profile's workflow_kind: support_inbox returns {open_threads, ack_lag, owner_gaps, escalations, next_actions}; incident_room returns {incident_summary, timeline, open_risks, owner_gaps, next_actions}; exec_brief returns {summary, decisions, risks, asks, action_items}; product_launch_watch returns {launch_signals, feedback_themes, blockers, metrics, next_actions}; custom returns {summary, highlights, open_questions, next_actions}. Hosted-only. Free tier ships 3 calls/month; Pro $9/mo unlocks unlimited (scheduled morning DM at 8am workspace tz rolling out Q2 2026).",
+    description: "Run a structured catch-up against a saved workflow profile. Returns structured JSON per the profile's workflow_kind: support_inbox returns {open_threads, ack_lag, owner_gaps, escalations, next_actions}; incident_room returns {incident_summary, timeline, open_risks, owner_gaps, next_actions}; exec_brief returns {summary, decisions, risks, asks, action_items}; product_launch_watch returns {launch_signals, feedback_themes, blockers, metrics, next_actions}; custom returns {summary, highlights, open_questions, next_actions}. Hosted-only. Free tier ships 25 AI tool calls/month (shared across the hosted AI tools); Pro $19/mo unlocks unlimited.",
     inputSchema: {
       type: "object",
       properties: {
@@ -561,7 +561,7 @@ export const TOOLS = [
   },
   {
     name: "slack_triage",
-    description: "Classify and route Slack threads against a workflow profile. Returns triage decisions per thread: priority (low|medium|high|urgent), suggested owner, escalation flag, time-sensitivity, and a routing recommendation. Hosted-only. Free tier ships 5 triage runs per day; Pro $9/mo unlocks unlimited.",
+    description: "Classify and route Slack threads against a workflow profile. Returns triage decisions per thread: priority (low|medium|high|urgent), suggested owner, escalation flag, time-sensitivity, and a routing recommendation. Hosted-only. Free tier ships 25 AI tool calls/month (shared across the hosted AI tools); Pro $19/mo unlocks unlimited.",
     inputSchema: {
       type: "object",
       properties: {

package/lib/workflow-store.js

@@ -70,32 +70,47 @@ export function loadStore() {
   try {
     const data = JSON.parse(raw);
     if (!data || typeof data !== "object" || !data.profiles) {
-      backupCorruptStore(raw, "shape-invalid");
+      quarantineCorruptStore("shape-invalid");
       return emptyStore();
     }
     if (data.version !== STORE_VERSION) {
-      // Future: migration logic. For now, back up the unrecognized version
-      // before falling back to empty so the old data is recoverable.
-      backupCorruptStore(raw, `version-${data.version}`);
+      // Future: migration logic. For now, quarantine the unrecognized
+      // version before falling back to empty so the old data is recoverable.
+      quarantineCorruptStore(`version-${data.version}`);
       return emptyStore();
     }
     return data;
   } catch {
-    backupCorruptStore(raw, "json-parse-error");
+    quarantineCorruptStore("json-parse-error");
     return emptyStore();
   }
 }
 
-function backupCorruptStore(raw, reasonTag) {
+/**
+ * Move an unreadable store aside to <store>.corrupt-<timestamp> before
+ * loadStore falls back to an empty store. MOVING (not copying) matters:
+ * it guarantees the next saveStore cannot clobber the user's only copy of
+ * their profiles, and repeated loads don't spawn one backup per call.
+ * Warns on stderr (stdout belongs to the MCP stdio protocol).
+ */
+function quarantineCorruptStore(reasonTag) {
+  const stamp = new Date().toISOString().replace(/[:.]/g, "-");
+  const quarantinePath = `${STORE_FILE}.corrupt-${stamp}`;
   try {
-    const stamp = new Date().toISOString().replace(/[:.]/g, "-");
-    const backupPath = `${STORE_FILE}.bak.${reasonTag}.${stamp}`;
-    writeFileSync(backupPath, raw);
+    renameSync(STORE_FILE, quarantinePath);
     if (platform() === "darwin" || platform() === "linux") {
-      try { chmodSync(backupPath, 0o600); } catch {}
+      try { chmodSync(quarantinePath, 0o600); } catch {}
     }
-  } catch {
-    // Backup is best-effort; do not throw on backup failure.
+    console.error(
+      `[slack-mcp] WARNING: workflow store ${STORE_FILE} is unreadable (${reasonTag}). ` +
+      `Quarantined it to ${quarantinePath} and continuing with an empty store.`
+    );
+  } catch (e) {
+    // Quarantine is best-effort; never let it break loadStore.
+    console.error(
+      `[slack-mcp] WARNING: workflow store ${STORE_FILE} is unreadable (${reasonTag}) ` +
+      `and could not be quarantined: ${e?.message || e}`
+    );
   }
 }
 

package/package.json

@@ -1,8 +1,8 @@
 {
   "name": "@jtalk22/slack-mcp",
   "mcpName": "io.github.jtalk22/slack-mcp-server",
-  "version": "4.4.1",
-  "description": "Slack MCP without OAuth \u2014 21 tools, session-based, local-first. Free OSS + hosted tier from $9/mo.",
+  "version": "4.4.3",
+  "description": "Slack MCP without OAuth — 21 tools, session-based, local-first. Free OSS + hosted tier from $19/mo.",
   "type": "module",
   "main": "src/server.js",
   "bin": {

package/public/share.html

@@ -124,7 +124,7 @@
       <a href="https://mcp.revasserlabs.com" rel="noopener" style="background:rgba(240,194,70,0.18);border-color:rgba(240,194,70,0.45);color:#f0c246">Hosted</a>
     </div>
 
-    <p class="note"><strong>Verify in 30 seconds:</strong> <code>--version</code>, <code>--doctor</code>, <code>--status</code>. Self-host gives 21 tools with session-based auth. Works with any MCP client — Claude, ChatGPT, Cursor, Copilot, Gemini, Windsurf. Hosted free tier (no card) live at <a href="https://mcp.revasserlabs.com">mcp.revasserlabs.com</a> — Pro $9/mo unlocks unlimited AI tools (scheduled morning catch-up DM rolling out Q2 2026).</p>
+    <p class="note"><strong>Verify in 30 seconds:</strong> <code>--version</code>, <code>--doctor</code>, <code>--status</code>. Self-host gives 21 tools with session-based auth. Works with any MCP client — Claude, ChatGPT, Cursor, Copilot, Gemini, Windsurf. Hosted free tier (no card) live at <a href="https://mcp.revasserlabs.com">mcp.revasserlabs.com</a> — Pro $19/mo unlocks unlimited AI tools and permanent OAuth.</p>
   </main>
 </body>
 </html>

package/scripts/apply-template.js

@@ -43,7 +43,7 @@ function printUsage() {
   console.log("  slack-mcp --apply-template support-triage --channels C012345,C067890");
   console.log("");
   console.log("Templates write to ~/.slack-mcp-workflows.json. The hosted AI brain at");
-  console.log("mcp.revasserlabs.com (free tier or Pro $9/mo) reads these profiles and");
+  console.log("mcp.revasserlabs.com (free tier or Pro $19/mo) reads these profiles and");
   console.log("returns structured JSON per the workflow_kind. The OSS package ships the");
   console.log("profile primitives + 3 discoverable upgrade stubs (slack_smart_search,");
   console.log("slack_catch_me_up, slack_triage). The brain is hosted-only.");
@@ -113,5 +113,5 @@ if (!profile.channels.length) {
   console.log("Or call slack_workflow_save from your MCP client to update.");
 } else {
   console.log("Profile is ready. Run slack_catch_me_up against it from your MCP client.");
-  console.log(`(Free tier: 3 catch_me_up calls/month. Pro $9/mo unlocks unlimited; scheduled morning DM rolling out Q2 2026.)`);
+  console.log(`(Free tier: 25 AI tool calls/month. Pro $19/mo unlocks unlimited.)`);
 }

package/scripts/setup-wizard.js

@@ -474,7 +474,7 @@ async function showHelp() {
   print("  https://github.com/jtalk22/slack-mcp-server");
   print();
   print(`${colors.bold}Hosted tier:${colors.reset}`);
-  print("  https://mcp.revasserlabs.com — $9/mo Pro, permanent OAuth,");
+  print("  https://mcp.revasserlabs.com — $19/mo Pro, permanent OAuth,");
   print("  semantic search, workflow continuity across channels.");
 }
 
@@ -548,7 +548,7 @@ async function main() {
       print("  • Or add to Claude Desktop config");
       print();
       print(`${colors.dim}Want permanent tokens, semantic search, and workflow continuity?${colors.reset}`);
-      print(`${colors.dim}Hosted tier: https://mcp.revasserlabs.com — $9/mo Pro, 10 free paid calls.${colors.reset}`);
+      print(`${colors.dim}Hosted tier: https://mcp.revasserlabs.com — $19/mo Pro, 25 free AI calls/mo.${colors.reset}`);
     } else {
       print(`${colors.red}Setup failed.${colors.reset} See errors above.`);
       process.exit(1);

package/server.json

@@ -2,7 +2,7 @@
   "$schema": "https://static.modelcontextprotocol.io/schemas/2025-12-11/server.schema.json",
   "name": "io.github.jtalk22/slack-mcp-server",
   "title": "Slack MCP Server",
-  "description": "Slack MCP without OAuth \u2014 21 tools, session-based, local-first. Free OSS + hosted tier from $9/mo.",
+  "description": "Slack MCP without OAuth \u2014 21 tools, session-based, local-first. Free OSS + hosted tier from $19/mo.",
   "websiteUrl": "https://mcp.revasserlabs.com",
   "icons": [
     {
@@ -17,7 +17,7 @@
     "url": "https://github.com/jtalk22/slack-mcp-server",
     "source": "github"
   },
-  "version": "4.4.1",
+  "version": "4.4.3",
   "remotes": [
     {
       "type": "streamable-http",
@@ -28,7 +28,7 @@
     {
       "registryType": "npm",
       "identifier": "@jtalk22/slack-mcp",
-      "version": "4.4.1",
+      "version": "4.4.3",
       "transport": {
         "type": "stdio"
       },

package/src/server-http.js

@@ -15,25 +15,9 @@ import {
 } from "@modelcontextprotocol/sdk/types.js";
 
 import { TOOLS } from "../lib/tools.js";
-import {
-  handleTokenStatus,
-  handleHealthCheck,
-  handleRefreshTokens,
-  handleListConversations,
-  handleConversationsHistory,
-  handleGetFullConversation,
-  handleSearchMessages,
-  handleUsersInfo,
-  handleSendMessage,
-  handleGetThread,
-  handleListUsers,
-  handleAddReaction,
-  handleRemoveReaction,
-  handleConversationsMark,
-  handleConversationsUnreads,
-  handleUsersSearch,
-} from "../lib/handlers.js";
+import { TOOL_HANDLERS } from "../lib/handlers.js";
 import { RELEASE_VERSION } from "../lib/public-metadata.js";
+import { isAuthDeath, lifeboatResponse } from "../lib/lifeboat.js";
 
 const SERVER_NAME = "slack-mcp-server";
 const SERVER_VERSION = RELEASE_VERSION;
@@ -98,54 +82,29 @@ server.setRequestHandler(CallToolRequestSchema, async (request) => {
   const { name, arguments: args } = request.params;
 
   try {
-    switch (name) {
-      case "slack_token_status":
-        return await handleTokenStatus();
-      case "slack_health_check":
-        return await handleHealthCheck();
-      case "slack_refresh_tokens":
-        return await handleRefreshTokens();
-      case "slack_list_conversations":
-        return await handleListConversations(args);
-      case "slack_conversations_history":
-        return await handleConversationsHistory(args);
-      case "slack_get_full_conversation":
-        return await handleGetFullConversation(args);
-      case "slack_search_messages":
-        return await handleSearchMessages(args);
-      case "slack_users_info":
-        return await handleUsersInfo(args);
-      case "slack_send_message":
-        return await handleSendMessage(args);
-      case "slack_get_thread":
-        return await handleGetThread(args);
-      case "slack_list_users":
-        return await handleListUsers(args);
-      case "slack_add_reaction":
-        return await handleAddReaction(args);
-      case "slack_remove_reaction":
-        return await handleRemoveReaction(args);
-      case "slack_conversations_mark":
-        return await handleConversationsMark(args);
-      case "slack_conversations_unreads":
-        return await handleConversationsUnreads(args);
-      case "slack_users_search":
-        return await handleUsersSearch(args);
-      default:
-        return {
-          content: [{
-            type: "text",
-            text: JSON.stringify({
-              status: "error",
-              code: "unknown_tool",
-              message: `Unknown tool: ${name}`,
-              next_action: "Call tools/list to inspect available tool names."
-            }, null, 2)
-          }],
-          isError: true
-        };
+    // Shared dispatch map keeps this transport in lockstep with the stdio
+    // server and the advertised TOOLS list.
+    const handler = TOOL_HANDLERS[name];
+    if (!handler) {
+      return {
+        content: [{
+          type: "text",
+          text: JSON.stringify({
+            status: "error",
+            code: "unknown_tool",
+            message: `Unknown tool: ${name}`,
+            next_action: "Call tools/list to inspect available tool names."
+          }, null, 2)
+        }],
+        isError: true
+      };
     }
+    return await handler(args);
   } catch (error) {
+    // OAuth Lifeboat: dead session token → recovery guidance, not a raw error.
+    if (isAuthDeath(error)) {
+      return lifeboatResponse(error);
+    }
     return {
       content: [{
         type: "text",

package/src/server.js

@@ -30,28 +30,11 @@ import { RELEASE_VERSION } from "../lib/public-metadata.js";
 import { checkTokenHealth } from "../lib/slack-client.js";
 import { TOOLS } from "../lib/tools.js";
 import {
-  handleTokenStatus,
+  TOOL_HANDLERS,
   handleHealthCheck,
-  handleRefreshTokens,
   handleListConversations,
-  handleConversationsHistory,
-  handleGetFullConversation,
-  handleSearchMessages,
-  handleUsersInfo,
-  handleSendMessage,
-  handleGetThread,
-  handleListUsers,
-  handleAddReaction,
-  handleRemoveReaction,
-  handleConversationsMark,
-  handleConversationsUnreads,
-  handleUsersSearch,
-  handleWorkflowSave,
-  handleWorkflows,
-  handleSmartSearch,
-  handleCatchMeUp,
-  handleTriage,
 } from "../lib/handlers.js";
+import { isAuthDeath, lifeboatResponse } from "../lib/lifeboat.js";
 
 // Background refresh interval (4 hours)
 const BACKGROUND_REFRESH_INTERVAL = 4 * 60 * 60 * 1000;
@@ -231,103 +214,29 @@ server.setRequestHandler(CallToolRequestSchema, async (request) => {
   const { name, arguments: args } = request.params;
 
   try {
-    switch (name) {
-      case "slack_token_status":
-        return await handleTokenStatus();
-
-      case "slack_health_check":
-        return await handleHealthCheck();
-
-      case "slack_refresh_tokens":
-        return await handleRefreshTokens();
-
-      case "slack_list_conversations":
-        return await handleListConversations(args);
-
-      case "slack_conversations_history":
-        return await handleConversationsHistory(args);
-
-      case "slack_get_full_conversation":
-        return await handleGetFullConversation(args);
-
-      case "slack_search_messages":
-        return await handleSearchMessages(args);
-
-      case "slack_users_info":
-        return await handleUsersInfo(args);
-
-      case "slack_send_message":
-        return await handleSendMessage(args);
-
-      case "slack_get_thread":
-        return await handleGetThread(args);
-
-      case "slack_list_users":
-        return await handleListUsers(args);
-
-      case "slack_add_reaction":
-        return await handleAddReaction(args);
-
-      case "slack_remove_reaction":
-        return await handleRemoveReaction(args);
-
-      case "slack_conversations_mark":
-        return await handleConversationsMark(args);
-
-      case "slack_conversations_unreads":
-        return await handleConversationsUnreads(args);
-
-      case "slack_users_search":
-        return await handleUsersSearch(args);
-
-      // Workflow profile primitives (OSS local JSON store)
-      case "slack_workflow_save":
-        return await handleWorkflowSave(args);
-
-      case "slack_workflows":
-        return await handleWorkflows(args);
-
-      // Hosted-only AI tools (OSS = upgrade stubs)
-      case "slack_smart_search":
-        return await handleSmartSearch(args);
-
-      case "slack_catch_me_up":
-        return await handleCatchMeUp(args);
-
-      case "slack_triage":
-        return await handleTriage(args);
-
-      default:
-        return {
-          content: [{
-            type: "text",
-            text: JSON.stringify({
-              status: "error",
-              code: "unknown_tool",
-              message: `Unknown tool: ${name}`,
-              next_action: "Use tools/list to discover available tools."
-            }, null, 2)
-          }],
-          isError: true
-        };
-    }
-  } catch (error) {
-    if (error?.code === "token_auth_failed") {
+    // Shared dispatch map (lib/handlers.js) keeps this transport in lockstep
+    // with the HTTP server and the advertised TOOLS list.
+    const handler = TOOL_HANDLERS[name];
+    if (!handler) {
       return {
         content: [{
           type: "text",
           text: JSON.stringify({
             status: "error",
-            code: "token_auth_failed",
-            message: String(error?.message || error),
-            slack_error: error.slack_error || null,
-            extraction_error: error.extraction_error || null,
-            next_action: error.next_action || "Open http://localhost:3000 and click Refresh, OR run `npm run tokens:auto` with Slack open in Chrome, OR check Chrome > View > Developer > Allow JavaScript from Apple Events."
+            code: "unknown_tool",
+            message: `Unknown tool: ${name}`,
+            next_action: "Use tools/list to discover available tools."
           }, null, 2)
         }],
         isError: true
       };
     }
+    return await handler(args);
+  } catch (error) {
+    // OAuth Lifeboat: dead session token → recovery guidance, not a raw error.
+    if (isAuthDeath(error)) {
+      return lifeboatResponse(error);
+    }
     return {
       content: [{
         type: "text",