tachibot-mcp 2.26.1 → 2.27.0

package/README.md

@@ -4,18 +4,18 @@
 
 ### Multi-Model AI Orchestration Platform
 
-[![Version](https://img.shields.io/badge/version-2.26.1-blue.svg)](https://www.npmjs.com/package/tachibot-mcp)
-[![Tools](https://img.shields.io/badge/tools-64_active-brightgreen.svg)](#-tool-ecosystem-64-tools)
+[![Version](https://img.shields.io/badge/version-2.27.0-blue.svg)](https://www.npmjs.com/package/tachibot-mcp)
+[![Tools](https://img.shields.io/badge/tools-65_active-brightgreen.svg)](#-tool-ecosystem-65-tools)
 [![License](https://img.shields.io/badge/license-AGPL--3.0-green.svg)](LICENSE)
 [![Node](https://img.shields.io/badge/node-%3E%3D22.0.0-brightgreen.svg)](https://nodejs.org)
 [![MCP](https://img.shields.io/badge/MCP-Compatible-purple.svg)](https://modelcontextprotocol.io)
 
-**64 AI tools. 12 providers. One protocol.**
+**65 AI tools. 12 providers. One protocol.**
 
 Orchestrate Perplexity, Grok, GPT-5.5, Gemini, Qwen, Kimi K2.7-Code, and MiniMax M3
 from Claude Code, Claude Desktop, Cursor, or any MCP client.
 
-[Get Started](#-quick-start) · [View Tools](#-tool-ecosystem-64-tools) · [Documentation](https://tachibot.com/docs)
+[Get Started](#-quick-start) · [View Tools](#-tool-ecosystem-65-tools) · [Documentation](https://tachibot.com/docs)
 
 <br>
 
@@ -30,6 +30,11 @@ from Claude Code, Claude Desktop, Cursor, or any MCP client.
 
 ## What's New
 
+### v2.27.0
+- **Grok 4.5** (Jul 8 flagship, "Opus-class") now powers all Grok tools, with automatic fallback to `grok-4.3` while xAI's region-staged rollout completes (EU mid-July) — tools keep working everywhere, and 4.5 activates by itself.
+- **`grok_search_lite`** (new tool, 65 total) — the same Grok live search on `grok-4-1-fast` ($0.20/$0.50, 2M ctx), ~10x cheaper than `grok_search`. Use it for high-volume lookups and jury/council fan-outs.
+- **GPT-5.6 tiers** — `openai_*` tools move to `gpt-5.6-sol` (flagship, same $5/$30 as 5.5 but stronger), `terra` for code (5.5-level at half price), `luna` for explanations ($1/$6). The $30/$180 `gpt-5.5-pro` tier is replaced by sol + reasoning effort; a permission fallback (sol → terra → 5.5) covers org-gated accounts.
+
 ### v2.26.1
 - **`/test` and `/audit` skills** (19 skills total) — `/test` generates runnable tests via `testgen`; `/audit` runs an OWASP/CWE security review via `security_review`.
 - **Skill install in the wizard** — `tachibot init` now offers to install Claude Code skills with a per-skill skip choice (`[Enter]`=all · `[s]`=choose which to skip · `[n]`=none). Skills are opt-in — `postinstall` no longer writes to `~/.claude` silently (`npm run install-skills` still installs all non-interactively).
@@ -47,7 +52,7 @@ from Claude Code, Claude Desktop, Cursor, or any MCP client.
 - **One-click Claude Desktop install** &mdash; download the `.mcpb` from the latest release and double-click. No JSON editing.
 - **`doctor`** &mdash; shows which keys are set, which tools are visible vs hidden and why, and what to try first.
 
-#### New tools & skills (64 tools · 19 skills)
+#### New tools & skills (65 tools · 19 skills)
 - `debug_triage` &mdash; ranked root-cause hypotheses with the cheapest discriminating check for each (Grok 4.3)
 - `spec_writer` &mdash; loose request → reviewable spec: user stories, Given/When/Then, out-of-scope, open questions (GPT-5.5)
 - `diff_review` / `plan_critique` / `testgen` / `security_review` &mdash; multi-model diff review, adversarial plan red-team, test generation, OWASP/CWE audit
@@ -95,7 +100,7 @@ Skills automatically adapt to your configured API keys. Even with just 1-2 provi
 ## Key Features
 
 ### Multi-Model Intelligence
-- **64 AI Tools** across 12 providers &mdash; Perplexity, Grok, GPT-5, Gemini, Qwen, Kimi, MiniMax, DeepSeek, GLM (Zhipu), StepFun, ERNIE (Baidu), plus free local models (Ollama / LM Studio / llama.cpp / vLLM)
+- **65 AI Tools** across 12 providers &mdash; Perplexity, Grok, GPT-5, Gemini, Qwen, Kimi, MiniMax, DeepSeek, GLM (Zhipu), StepFun, ERNIE (Baidu), plus free local models (Ollama / LM Studio / llama.cpp / vLLM)
 - **Gemini 3.5 Flash** (`gemini-3.5-flash`, GA May 19 2026) &mdash; Flash/search tier; reasoning default stays `gemini-3.1-pro-preview`
 - **Multi-Model Council** &mdash; planner_maker synthesizes plans from 5+ models into bite-sized TDD steps
 - **Smart Routing** &mdash; Automatic model selection for optimal results
@@ -111,11 +116,11 @@ Skills automatically adapt to your configured API keys. Even with just 1-2 provi
 | Profile | Tools | Best For |
 |---------|-------|----------|
 | **Minimal** | 13 | Quick tasks, low token budget |
-| **Research Power** | 35 | Deep investigation, multi-source |
+| **Research Power** | 36 | Deep investigation, multi-source |
 | **Code Focus** | 42 | Software development, SWE tasks |
-| **Balanced** | 53 | General-purpose, mixed workflows |
-| **Heavy Coding** | 57 | Max code tools + agentic workflows |
-| **Full** (default) | 64 | Everything enabled |
+| **Balanced** | 54 | General-purpose, mixed workflows |
+| **Heavy Coding** | 58 | Max code tools + agentic workflows |
+| **Full** (default) | 65 | Everything enabled |
 
 ### Developer Experience
 - **Claude Code** &mdash; First-class support
@@ -195,10 +200,10 @@ See [Installation Guide](docs/INSTALLATION_BOTH.md) for detailed instructions.
 
 ---
 
-## Tool Ecosystem (64 Tools)
+## Tool Ecosystem (65 Tools)
 
-### Research & Search (5)
-`perplexity_ask` &#183; `perplexity_reason` &#183; `grok_search` &#183; `openai_search` &#183; `gemini_search`
+### Research & Search (6)
+`perplexity_ask` &#183; `perplexity_reason` &#183; `grok_search` &#183; `grok_search_lite` &#183; `openai_search` &#183; `gemini_search`
 
 ### Reasoning & Planning (14)
 `grok_reason` &#183; `openai_reason` &#183; `qwen_reason` &#183; `qwq_reason` &#183; `kimi_thinking` &#183; `kimi_decompose` &#183; `deepseek_reason` &#183; `glm_reason` &#183; `stepfun_reason` &#183; `ernie_reason` &#183; `planner_maker` &#183; `planner_runner` &#183; `list_plans` &#183; `spec_writer`

package/dist/src/config/model-constants.js

@@ -6,26 +6,33 @@
 // =============================================================================
 // OPENAI MODELS (provider-based naming)
 // =============================================================================
-// GPT-5.5 released Apr 23, 2026 - CURRENT (agentic-focused, 1.1M context, omnimodal)
-// GPT-5.4 released Mar 5, 2026 - previous flagship (kept for MINI tier — no 5.5-mini yet)
-// Model is "gpt-5.5", "thinking" is controlled by reasoning.effort parameter
-// OpenRouter uses prefix: openai/gpt-5.5, openai/gpt-5.5-pro
+// GPT-5.6 released Jul 9, 2026 - CURRENT. Three durable tiers replace the pro/mini split:
+//   sol (flagship $5/$30) / terra (balanced $2.50/$15) / luna (fast $1/$6). All 1.05M ctx, 128K out.
+// "Pro" is no longer a model: sol + reasoning.effort (the $30/$180 gpt-5.5-pro tier is gone).
+// GPT-5.5 (Apr 23, 2026) kept as fallback. OpenRouter uses prefix: openai/gpt-5.6-sol etc.
 export const OPENAI_MODELS = {
-    // GPT-5.5 (Apr 23, 2026 - CURRENT)
-    // Note: "gpt-5.5" + reasoning.effort="high"/"xhigh" = "thinking" mode
-    DEFAULT: "gpt-5.5", // Most capable - agentic, reasoning, omnimodal (1.1M ctx)
-    MINI: "gpt-5.4-mini", // Fast/efficient coding & subagents (no 5.5-mini available yet)
-    PRO: "gpt-5.5-pro", // Expert: higher compute for harder problems ($30/$180)
+    // GPT-5.6 tiers (Jul 9, 2026 - CURRENT)
+    DEFAULT: "gpt-5.6-sol", // Flagship: coding/agentic SOTA, 54% more token-efficient than 5.5
+    SOL: "gpt-5.6-sol", // Explicit tier alias (also aliased "gpt-5.6" server-side)
+    TERRA: "gpt-5.6-terra", // Balanced: matches gpt-5.5 at half price ($2.50/$15)
+    LUNA: "gpt-5.6-luna", // Fast/cheap: $1/$6 (weaker long-context recall — avoid huge inputs)
+    MINI: "gpt-5.6-luna", // "mini" tier successor (was gpt-5.4-mini)
+    PRO: "gpt-5.6-sol", // Pro tier eliminated in 5.6 — sol + reasoning.effort xhigh (was gpt-5.5-pro $30/$180)
+    // GPT-5.5 (Apr 2026) - fallback
+    FALLBACK_5_5: "gpt-5.5",
     // Aliases for backward compatibility
-    THINKING: "gpt-5.5", // "Thinking" = gpt-5.5 with high reasoning effort
-    FULL: "gpt-5.5", // Map old FULL to DEFAULT
-    CODEX: "gpt-5.4-mini", // Map old codex to MINI (no 5.5 mini yet)
-    CODEX_MINI: "gpt-5.4-mini", // Map old codex-mini to MINI
-    CODEX_MAX: "gpt-5.5-pro", // Map old codex-max to PRO
-    INSTANT: "gpt-5.4-mini", // Map old instant to MINI
+    THINKING: "gpt-5.6-sol", // "Thinking" = sol with high reasoning effort
+    FULL: "gpt-5.6-sol", // Map old FULL to DEFAULT
+    CODEX: "gpt-5.6-terra", // Coding workhorse tier
+    CODEX_MINI: "gpt-5.6-luna", // Cheapest tier
+    CODEX_MAX: "gpt-5.6-sol", // Map old codex-max to flagship (use high effort)
+    INSTANT: "gpt-5.6-luna", // Fast tier
 };
 // OpenRouter model ID mapping (add prefix when using OpenRouter gateway)
 export const OPENROUTER_PREFIX_MAP = {
+    "gpt-5.6-sol": "openai/",
+    "gpt-5.6-terra": "openai/",
+    "gpt-5.6-luna": "openai/",
     "gpt-5.5": "openai/",
     "gpt-5.5-pro": "openai/",
     "gpt-5.4-mini": "openai/",
@@ -63,18 +70,21 @@ export const PERPLEXITY_MODELS = {
     SONAR_PRO: "sonar-pro", // Advanced search
     SONAR_REASONING: "sonar-reasoning-pro", // Reasoning model (expensive - avoid)
 };
-// Grok Models (xAI) - Updated 2026-06-01 with Grok 4.3 (Apr 30, 2026 flagship)
+// Grok Models (xAI) - Updated 2026-07-11 with Grok 4.5 (Jul 8, 2026 flagship)
 export const GROK_MODELS = {
-    // Grok 4.3 (Apr 30, 2026) - CURRENT FLAGSHIP
-    // Single model ID with configurable reasoning effort (replaces 4.20's reasoning/non-reasoning/multi-agent split).
-    // 1M context, $1.25/$2.50 (cheaper than 4.20), xAI's recommended model, lowest hallucination rate.
-    _4_3: "grok-4.3", // Flagship: 1M ctx, $1.25/$2.50, reasoning.effort low|high
+    // Grok 4.5 (Jul 8, 2026) - CURRENT FLAGSHIP ("Opus-class")
+    // 500K context (SMALLER than 4.3's 1M — fine for tools, mind huge inputs), $2/$6,
+    // configurable reasoning effort. EU API rollout was "mid-July" — 4.3 stays the fallback.
+    _4_5: "grok-4.5", // Flagship: 500K ctx, $2/$6, reasoning.effort low|high
+    // Grok 4.3 (Apr 30, 2026) - previous flagship, kept as FALLBACK
+    // 1M context, $1.25/$2.50, lowest verified hallucination rate.
+    _4_3: "grok-4.3", // Fallback: 1M ctx, $1.25/$2.50, reasoning.effort low|high
     _4_3_LATEST: "grok-4.3-latest", // Rolling alias for newest 4.3 snapshot
     _BUILD: "grok-build-0.1", // Coding specialist (May 29, 2026): 256k ctx, fast agentic coding
-    // Grok 4.20 models (Mar 10, 2026) - LEGACY (still valid; kept as fallback). Now resolve to 4.3 via deprecated keys.
-    _4_20_REASONING: "grok-4.3", // [deprecated key] → grok-4.3 (was grok-4.20-0309-reasoning)
-    _4_20_NON_REASONING: "grok-4.3", // [deprecated key] → grok-4.3 (was grok-4.20-0309-non-reasoning)
-    _4_20_MULTI_AGENT: "grok-4.3", // [deprecated key] → grok-4.3 (architect uses high reasoning.effort)
+    // Grok 4.20 models (Mar 10, 2026) - LEGACY keys. Deprecated keys now resolve to the current flagship (4.5).
+    _4_20_REASONING: "grok-4.5", // [deprecated key] → grok-4.5 (was grok-4.20-0309-reasoning)
+    _4_20_NON_REASONING: "grok-4.5", // [deprecated key] → grok-4.5 (was grok-4.20-0309-non-reasoning)
+    _4_20_MULTI_AGENT: "grok-4.5", // [deprecated key] → grok-4.5 (architect uses high reasoning.effort)
     // Grok 4.1 fast models (Nov 2025) - BEST VALUE (10x cheaper)
     _4_1_FAST_REASONING: "grok-4-1-fast-reasoning", // Fast reasoning: 2M context, $0.20/$0.50
     _4_1_FAST_NON_REASONING: "grok-4-1-fast-non-reasoning", // Tool-calling optimized: 2M context, $0.20/$0.50
@@ -159,26 +169,27 @@ export const DEFAULT_WORKFLOW_SETTINGS = {
 // When new models release, update ONLY this section!
 // All tools automatically use the new models.
 // ============================================================================
-// UPDATED Apr 26, 2026: GPT-5.5 (flagship, agentic-focused) + GPT-5.4-mini (coding/fast — no 5.5-mini yet)
-// Kimi K2.6 (Apr 20, 2026 - SWE-bench Pro leader)
+// UPDATED Jul 11, 2026: GPT-5.6 tiers (sol/terra/luna — Jul 9) + Grok 4.5 (Jul 8 flagship)
+// Kimi K2.7-Code (Jun 12, 2026)
 export const CURRENT_MODELS = {
     openai: {
-        default: OPENAI_MODELS.DEFAULT, // gpt-5.5 - most capable, agentic (Apr 2026)
-        reason: OPENAI_MODELS.DEFAULT, // Deep reasoning (gpt-5.5 + effort=high)
-        brainstorm: OPENAI_MODELS.DEFAULT, // Creative ideation (gpt-5.5 + effort=medium)
-        code: OPENAI_MODELS.MINI, // Code tasks (gpt-5.4-mini - cheapest tier, no 5.5-mini yet)
-        explain: OPENAI_MODELS.MINI, // Explanations (gpt-5.4-mini - fast & capable)
-        search: OPENAI_MODELS.DEFAULT, // Web search (gpt-5.5 + web_search tool)
-        // Premium option for opt-in (use sparingly - $30/$180 per 1M tokens)
-        premium: OPENAI_MODELS.PRO, // Expert mode (gpt-5.5-pro - higher compute)
+        default: OPENAI_MODELS.DEFAULT, // gpt-5.6-sol - flagship, agentic/coding SOTA (Jul 2026)
+        reason: OPENAI_MODELS.DEFAULT, // Deep reasoning (sol + effort=high)
+        brainstorm: OPENAI_MODELS.DEFAULT, // Creative ideation (sol + effort=medium)
+        code: OPENAI_MODELS.TERRA, // Code tasks (gpt-5.6-terra - 5.5-level at half price)
+        explain: OPENAI_MODELS.LUNA, // Explanations (gpt-5.6-luna - $1/$6, fast)
+        search: OPENAI_MODELS.DEFAULT, // Web search (sol + web_search tool)
+        // Premium: the 5.5-pro tier ($30/$180) is gone — sol + effort=xhigh replaces it at $5/$30
+        premium: OPENAI_MODELS.PRO, // Expert mode (gpt-5.6-sol - crank reasoning effort)
     },
     grok: {
-        reason: GROK_MODELS._4_3, // grok-4.3 (flagship, low hallucination, high reasoning effort)
-        code: GROK_MODELS._4_3, // grok-4.3 (flagship quality, tool-calling)
-        debug: GROK_MODELS._4_3, // grok-4.3 (low hallucination for debugging)
-        brainstorm: GROK_MODELS._4_3, // grok-4.3 (1M context)
-        search: GROK_MODELS._4_3, // grok-4.3 LOW HALLUCINATION - critical for search
-        architect: GROK_MODELS._4_3, // grok-4.3 with high reasoning.effort (agentic swarm behaviour)
+        reason: GROK_MODELS._4_5, // grok-4.5 (Jul 2026 flagship, high reasoning effort)
+        code: GROK_MODELS._4_5, // grok-4.5 (flagship quality, tool-calling)
+        debug: GROK_MODELS._4_5, // grok-4.5 (flagship debugging)
+        brainstorm: GROK_MODELS._4_5, // grok-4.5 (500K context)
+        search: GROK_MODELS._4_5, // grok-4.5 flagship synthesis over live search
+        search_lite: GROK_MODELS._4_1_FAST_NON_REASONING, // grok-4-1-fast: $0.20/$0.50, 2M ctx, tool-calling optimized — 10x cheaper search
+        architect: GROK_MODELS._4_5, // grok-4.5 with high reasoning.effort (agentic swarm behaviour)
     },
     gemini: {
         default: GEMINI_MODELS.GEMINI_3_PRO,
@@ -275,6 +286,11 @@ export const TOOL_DEFAULTS = {
         maxTokens: 3000,
         temperature: 0.7,
     },
+    grok_search_lite: {
+        model: CURRENT_MODELS.grok.search_lite,
+        maxTokens: 3000,
+        temperature: 0.3,
+    },
     grok_brainstorm: {
         model: CURRENT_MODELS.grok.brainstorm,
         maxTokens: 4000,
@@ -343,6 +359,10 @@ export const DEFAULT_WORKFLOW_TOOL = "openai_brainstorm";
 // Used in tool outputs, usage stats, logs - keeps display consistent
 export const MODEL_DISPLAY_NAMES = {
     // OpenAI
+    "gpt-5.6-sol": "gpt-5.6-sol",
+    "gpt-5.6-terra": "gpt-5.6-terra",
+    "gpt-5.6-luna": "gpt-5.6-luna",
+    "gpt-5.6": "gpt-5.6-sol",
     "gpt-5.5": "gpt-5.5",
     "gpt-5.5-pro": "gpt-5.5-pro",
     "gpt-5.4": "gpt-5.4",
@@ -354,6 +374,7 @@ export const MODEL_DISPLAY_NAMES = {
     "gemini-3-flash-preview": "gemini-3-flash",
     "gemini-3.1-flash-lite": "gemini-3.1-flash-lite",
     // Grok (xAI)
+    "grok-4.5": "grok-4.5",
     "grok-4.3": "grok-4.3",
     "grok-4.3-latest": "grok-4.3",
     "grok-build-0.1": "grok-build",
@@ -397,8 +418,11 @@ export function getModelDisplayName(modelId) {
 // Model pricing per 1K tokens (input/output average) for cost tracking
 export const MODEL_PRICING = {
     // OpenAI
-    "gpt-5.5": 0.0175, // ($5 + $30) / 2 / 1000 (Apr 23, 2026)
-    "gpt-5.5-pro": 0.105, // ($30 + $180) / 2 / 1000 (Apr 23, 2026)
+    "gpt-5.6-sol": 0.0175, // ($5 + $30) / 2 / 1000 (Jul 9, 2026)
+    "gpt-5.6-terra": 0.00875, // ($2.50 + $15) / 2 / 1000 (Jul 9, 2026)
+    "gpt-5.6-luna": 0.0035, // ($1 + $6) / 2 / 1000 (Jul 9, 2026)
+    "gpt-5.5": 0.0175, // ($5 + $30) / 2 / 1000 (Apr 23, 2026 - fallback)
+    "gpt-5.5-pro": 0.105, // ($30 + $180) / 2 / 1000 (Apr 23, 2026 - tier removed in 5.6)
     "gpt-5.4": 0.00875, // ($2.50 + $15) / 2 / 1000 (Mar 2026 - legacy)
     "gpt-5.4-mini": 0.002625, // ($0.75 + $4.50) / 2 / 1000 (Mar 17, 2026)
     "gpt-5.4-pro": 0.105, // ($30 + $180) / 2 / 1000 (Mar 2026)
@@ -408,7 +432,8 @@ export const MODEL_PRICING = {
     "gemini-3-flash-preview": 0.00175, // ($0.50 + $3) / 2 / 1000 (legacy)
     "gemini-3.1-flash-lite": 0.001, // Cheapest/fastest in 3.1 series (Mar 2026)
     // Grok
-    "grok-4.3": 0.001875, // ($1.25 + $2.50) / 2 / 1000 (Apr 30, 2026 - cheaper than 4.20)
+    "grok-4.5": 0.004, // ($2 + $6) / 2 / 1000 (Jul 8, 2026 flagship)
+    "grok-4.3": 0.001875, // ($1.25 + $2.50) / 2 / 1000 (Apr 30, 2026 - fallback)
     "grok-4.3-latest": 0.001875,
     "grok-build-0.1": 0.001875, // Coding specialist (estimate, same tier)
     "grok-4.20-0309-reasoning": 0.004, // ($2 + $6) / 2 / 1000 (legacy)

package/dist/src/profiles/balanced.js

@@ -1,5 +1,5 @@
 export const balancedProfile = {
-    description: "Balanced set for general use (53 tools)",
+    description: "Balanced set for general use (54 tools)",
     tools: {
         think: true,
         focus: true,
@@ -16,6 +16,7 @@ export const balancedProfile = {
         grok_architect: false,
         grok_brainstorm: false,
         grok_search: true,
+        grok_search_lite: true,
         openai_reason: true,
         openai_brainstorm: true,
         openai_code_review: false,

package/dist/src/profiles/code_focus.js

@@ -16,6 +16,7 @@ export const codeFocusProfile = {
         grok_architect: false,
         grok_brainstorm: false,
         grok_search: false,
+        grok_search_lite: false,
         openai_reason: false,
         openai_brainstorm: false,
         openai_code_review: true,

package/dist/src/profiles/full.js

@@ -1,5 +1,5 @@
 export const fullProfile = {
-    description: "Default profile — all tools enabled for maximum capability (64 tools)",
+    description: "Default profile — all tools enabled for maximum capability (65 tools)",
     tools: {
         think: true,
         focus: true,
@@ -16,6 +16,7 @@ export const fullProfile = {
         grok_architect: true,
         grok_brainstorm: true,
         grok_search: true,
+        grok_search_lite: true,
         openai_reason: true,
         openai_brainstorm: true,
         openai_code_review: true,

package/dist/src/profiles/heavy_coding.js

@@ -1,5 +1,5 @@
 export const heavyCodingProfile = {
-    description: "Heavy coding with all reasoning & code tools (57 tools)",
+    description: "Heavy coding with all reasoning & code tools (58 tools)",
     tools: {
         // Core reasoning - all enabled
         think: true,
@@ -19,6 +19,7 @@ export const heavyCodingProfile = {
         grok_architect: true,
         grok_brainstorm: true,
         grok_search: true,
+        grok_search_lite: true,
         // OpenAI - all enabled for heavy coding
         openai_reason: true,
         openai_brainstorm: true,

package/dist/src/profiles/minimal.js

@@ -16,6 +16,7 @@ export const minimalProfile = {
         grok_architect: false,
         grok_brainstorm: false,
         grok_search: false,
+        grok_search_lite: false,
         openai_reason: false,
         openai_brainstorm: false,
         openai_code_review: false,

package/dist/src/profiles/research_power.js

@@ -1,5 +1,5 @@
 export const researchPowerProfile = {
-    description: "Research-focused with Grok search + all Perplexity + brainstorming (35 tools)",
+    description: "Research-focused with Grok search + all Perplexity + brainstorming (36 tools)",
     tools: {
         think: true,
         focus: true,
@@ -16,6 +16,7 @@ export const researchPowerProfile = {
         grok_architect: false,
         grok_brainstorm: false,
         grok_search: true,
+        grok_search_lite: true,
         openai_reason: false,
         openai_brainstorm: true,
         openai_code_review: false,

package/dist/src/tools/grok-enhanced.js

@@ -38,9 +38,11 @@ export async function callGrokEnhanced(messages, options = {}) {
         // Use different endpoint and format for search vs non-search
         if (enableLiveSearch) {
             // NEW Agent Tools API (Jan 2025) - uses /v1/responses endpoint
-            // with 'input' instead of 'messages' and tools array
+            // with 'input' instead of 'messages' and tools array.
+            // Respects options.model so grok_search (flagship) and grok_search_lite
+            // (grok-4-1-fast, 10x cheaper) share this path.
             const searchRequestBody = {
-                model: GrokModel.GROK_4_20_NON_REASONING, // 4.20 standard is better for tool-calling search
+                model,
                 input: messages.map(m => ({ role: m.role, content: m.content })),
                 tools: [
                     { type: "web_search" },
@@ -132,8 +134,14 @@ export async function callGrokEnhanced(messages, options = {}) {
         };
     }
     catch (error) {
+        const errMsg = error instanceof Error ? error.message : String(error);
+        // grok-4.5's rollout is region-staged (EU lands mid-July 2026) — fall back to grok-4.3
+        if (model.startsWith('grok-4.5') && errMsg.includes('not available in your region')) {
+            console.error(`[Grok] grok-4.5 region-blocked, falling back to grok-4.3`);
+            return callGrokEnhanced(messages, { ...options, model: GrokModel.GROK_4_3 });
+        }
         return {
-            content: `[Grok error: ${error instanceof Error ? error.message : String(error)}]`
+            content: `[Grok error: ${errMsg}]`
         };
     }
 }
@@ -348,7 +356,7 @@ export const grokSearchTool = {
         const messages = [
             {
                 role: "system",
-                content: `You are Grok 4.3 with live search. Search for: "${query}".
+                content: `You are Grok 4.5 with live search. Search for: "${query}".
 ${recencyPrompt}
 Provide concise, factual results with sources.
 Limit search to ${max_search_results} sources for cost control.
@@ -359,13 +367,13 @@ ${FORMAT_INSTRUCTION}`
                 content: `Search for: ${query}`
             }
         ];
-        log?.info(`Grok Search: ${max_search_results} sources, recency: ${recency} (using grok-4.20 reasoning)`);
+        log?.info(`Grok Search: ${max_search_results} sources, recency: ${recency} (using grok-4.5 flagship)`);
         // Extract domains from sources if specified
         const domains = sources
             ?.filter((s) => s.allowed_websites)
             ?.flatMap((s) => s.allowed_websites) || [];
         const result = await callGrokEnhanced(messages, {
-            model: GrokModel.GROK_4_20_REASONING, // Low hallucination is CRITICAL for search accuracy
+            model: GrokModel.GROK_4_20_REASONING, // → grok-4.5 flagship (grok_search_lite is the cheap tier)
             enableLiveSearch: true,
             searchSources: max_search_results,
             searchDomains: domains,
@@ -403,11 +411,11 @@ export function isGrokAvailable() {
 export function getGrokStatus() {
     return {
         available: isGrokAvailable(),
-        model: "grok-4.3",
+        model: "grok-4.5",
         features: [
-            'Grok 4.3 (grok-4.3): Flagship, lowest hallucination, 1M context, configurable reasoning effort ($1.25/$2.50)',
+            'Grok 4.5 (grok-4.5): Flagship ("Opus-class", Jul 2026), 500K context, configurable reasoning effort ($2/$6)',
+            'Grok 4.3 (grok-4.3): Previous flagship, kept as fallback — 1M context, lowest verified hallucination ($1.25/$2.50)',
             'Grok Build (grok-build-0.1): Fast agentic coding specialist, 256k context',
-            'Grok 4.20 (grok-4.20-0309-*): Legacy flagship, still available as fallback (2M context)',
             'Live web search with citations',
             'Function calling',
             'Structured outputs',

package/dist/src/tools/grok-search-lite-tool.js

@@ -0,0 +1,82 @@
+/**
+ * Grok Search Lite — cost-optimized live web search.
+ *
+ * Same xAI live-search path as grok_search, but on grok-4-1-fast
+ * (tool-calling optimized, 2M context, $0.20/$0.50 — ~10x cheaper than the
+ * grok-4.5 flagship). Use for high-volume lookups, jury/council fan-outs, and
+ * quick fact checks; use grok_search when synthesis quality matters most.
+ */
+import { z } from "zod";
+import { callGrokEnhanced, GrokModel } from "./grok-enhanced.js";
+import { defineModelTool } from "./factory/define-model-tool.js";
+import { link } from "../utils/ansi-renderer.js";
+import { stripFormatting } from "../utils/format-stripper.js";
+import { FORMAT_INSTRUCTION } from "../utils/format-constants.js";
+/** Pure prompt builder — exported for tests. */
+export function buildGrokSearchLitePrompt(query, maxSearchResults, recency) {
+    const recencyPrompt = recency !== "all" ? `Focus on information from the last ${recency}.` : "";
+    return `You are Grok with live search (fast tier). Search for: "${query}".
+${recencyPrompt}
+Provide concise, factual results with sources.
+Limit search to ${maxSearchResults} sources for cost control.
+${FORMAT_INSTRUCTION}`;
+}
+export const grokSearchLiteTool = defineModelTool({
+    name: "grok_search_lite",
+    description: "Cheap web search (grok-4-1-fast, ~10x cheaper than grok_search). Put your QUERY in the 'query' parameter.",
+    parameters: z.object({
+        query: z.string(),
+        sources: z
+            .array(z.object({
+            type: z.enum(["web", "news", "x", "rss"]),
+            country: z.string().optional(),
+            allowed_websites: z.array(z.string()).optional(),
+        }))
+            .optional(),
+        max_search_results: z.number().optional(),
+        recency: z.enum(["all", "day", "week", "month", "year"]).optional(),
+    }),
+    execute: async (args, { log }) => {
+        const { query, sources = [{ type: "web" }], max_search_results = 20, recency = "all", } = args;
+        const messages = [
+            {
+                role: "system",
+                content: buildGrokSearchLitePrompt(query, max_search_results, recency),
+            },
+            {
+                role: "user",
+                content: `Search for: ${query}`,
+            },
+        ];
+        log?.info(`Grok Search Lite: ${max_search_results} sources, recency: ${recency} (using grok-4-1-fast, $0.20/$0.50)`);
+        // Extract domains from sources if specified
+        const domains = sources
+            ?.filter((s) => s.allowed_websites)
+            ?.flatMap((s) => s.allowed_websites) || [];
+        const result = await callGrokEnhanced(messages, {
+            model: GrokModel.GROK_4_1_FAST, // Tool-calling optimized cheap tier — the whole point of this tool
+            enableLiveSearch: true,
+            searchSources: max_search_results,
+            searchDomains: domains,
+            temperature: 0.3, // Low temperature for factual search
+            maxTokens: 3000,
+        });
+        // Format sources with ANSI links (same shape as grok_search output)
+        let output = result.content;
+        if (result.sources && result.sources.length > 0) {
+            const sourcesText = result.sources
+                .slice(0, 10)
+                .map((s, i) => {
+                const title = s.title || "Source";
+                const url = s.url || "";
+                return `  ${link(url, `[${i + 1}] ${title}`)}`;
+            })
+                .join("\n");
+            output += `\n\nSources:\n${sourcesText}`;
+        }
+        // Add cost info
+        const estimatedCost = (max_search_results / 1000) * 0.025;
+        output += `\n\nSearch used up to ${max_search_results} sources (~$${estimatedCost.toFixed(4)})`;
+        return stripFormatting(output);
+    },
+});

package/dist/src/tools/grok-tools.js

@@ -23,18 +23,20 @@ config({ path: path.resolve(__dirname, '../../../.env') });
 // Grok API configuration
 const GROK_API_KEY = getGrokApiKey();
 const GROK_API_URL = "https://api.x.ai/v1/chat/completions";
-// Available Grok models - Updated 2026-06-01 with Grok 4.3 (Apr 30, 2026 flagship)
+// Available Grok models - Updated 2026-07-11 with Grok 4.5 (Jul 8, 2026 flagship)
 export var GrokModel;
 (function (GrokModel) {
-    // Grok 4.3 (Apr 30, 2026) - CURRENT FLAGSHIP
-    // Single model ID with configurable reasoning effort; 1M ctx, $1.25/$2.50 (cheaper than 4.20).
+    // Grok 4.5 (Jul 8, 2026) - CURRENT FLAGSHIP ("Opus-class")
+    // 500K ctx (smaller than 4.3's 1M), $2/$6, configurable reasoning effort.
+    GrokModel["GROK_4_5"] = "grok-4.5";
+    // Grok 4.3 (Apr 30, 2026) - previous flagship, kept as FALLBACK
     GrokModel["GROK_4_3"] = "grok-4.3";
     GrokModel["GROK_4_3_LATEST"] = "grok-4.3-latest";
     GrokModel["GROK_BUILD"] = "grok-build-0.1";
-    // Grok 4.20 keys (Mar 2026) - DEPRECATED, retained for back-compat → now resolve to grok-4.3
-    GrokModel["GROK_4_20_REASONING"] = "grok-4.3";
-    GrokModel["GROK_4_20_NON_REASONING"] = "grok-4.3";
-    GrokModel["GROK_4_20_MULTI_AGENT"] = "grok-4.3";
+    // Grok 4.20 keys (Mar 2026) - DEPRECATED, retained for back-compat → now resolve to the current flagship
+    GrokModel["GROK_4_20_REASONING"] = "grok-4.5";
+    GrokModel["GROK_4_20_NON_REASONING"] = "grok-4.5";
+    GrokModel["GROK_4_20_MULTI_AGENT"] = "grok-4.5";
     // Grok 4.1 fast models (Nov 2025) - BEST VALUE (10x cheaper)
     GrokModel["GROK_4_1_FAST_REASONING"] = "grok-4-1-fast-reasoning";
     GrokModel["GROK_4_1_FAST"] = "grok-4-1-fast-non-reasoning";
@@ -53,7 +55,7 @@ export var GrokModel;
  *   - 'code-analysis': Relaxed for code analysis tools
  *   - 'llm-orchestration': Medium for LLM-to-LLM calls
  */
-export async function callGrok(messages, model = GrokModel.GROK_4_3, temperature = 0.7, maxTokens = 16384, // Increased default for comprehensive responses
+export async function callGrok(messages, model = GrokModel.GROK_4_5, temperature = 0.7, maxTokens = 16384, // Increased default for comprehensive responses
 forceVisibleOutput = true, validationContext = 'llm-orchestration', reasoningEffort) {
     // Try OpenRouter gateway first if enabled
     if (isGatewayEnabled()) {
@@ -78,9 +80,9 @@ forceVisibleOutput = true, validationContext = 'llm-orchestration', reasoningEff
         }
         return { ...msg, content: validation.sanitized };
     });
-    // Grok 4.x reasoning can take 60-90s; 4.20/4.3 flagship and multi-agent can take longer
+    // Grok 4.x reasoning can take 60-90s; 4.3/4.5 flagship and multi-agent can take longer
     const isReasoning = model.includes('reasoning') || model.includes('multi-agent');
-    const isFlagship = model.includes('4.20') || model.startsWith('grok-4.3');
+    const isFlagship = model.includes('4.20') || model.startsWith('grok-4.3') || model.startsWith('grok-4.5');
     const timeoutMs = isFlagship ? 180000 : (isReasoning ? 120000 : 60000);
     try {
         // For Grok 4+ models, we need to handle reasoning tokens specially
@@ -94,8 +96,8 @@ forceVisibleOutput = true, validationContext = 'llm-orchestration', reasoningEff
         }
         const controller = new AbortController();
         const timeoutId = setTimeout(() => controller.abort(), timeoutMs);
-        // grok-4.3 (single flagship) and 4.20 multi-agent both accept reasoning.effort
-        const supportsReasoningEffort = model.includes('multi-agent') || model.startsWith('grok-4.3');
+        // grok-4.3/4.5 flagships and 4.20 multi-agent all accept reasoning.effort
+        const supportsReasoningEffort = model.includes('multi-agent') || model.startsWith('grok-4.3') || model.startsWith('grok-4.5');
         const requestBody = {
             model,
             messages: validatedMessages,
@@ -138,7 +140,13 @@ forceVisibleOutput = true, validationContext = 'llm-orchestration', reasoningEff
         if (error instanceof Error && error.name === 'AbortError') {
             return `[Grok timeout: ${model} exceeded ${isReasoning ? '120' : '60'}s limit]`;
         }
-        return `[Grok error: ${error instanceof Error ? error.message : String(error)}]`;
+        const errMsg = error instanceof Error ? error.message : String(error);
+        // grok-4.5's rollout is region-staged (EU lands mid-July 2026) — fall back to grok-4.3
+        if (model.startsWith('grok-4.5') && errMsg.includes('not available in your region')) {
+            console.error(`[Grok] grok-4.5 region-blocked, falling back to grok-4.3`);
+            return callGrok(validatedMessages, GrokModel.GROK_4_3, temperature, maxTokens, forceVisibleOutput, validationContext, reasoningEffort);
+        }
+        return `[Grok error: ${errMsg}]`;
     }
 }
 /**
@@ -171,7 +179,7 @@ export const grokReasonTool = defineModelTool({
         const messages = [
             {
                 role: "system",
-                content: `You are Grok 4.3, an expert at logical reasoning and problem-solving.
+                content: `You are Grok 4.5, an expert at logical reasoning and problem-solving.
 ${approachPrompts[approach]}.
 ${context ? `Context: ${context}` : ''}
 ${FORMAT_INSTRUCTION}`
@@ -221,7 +229,7 @@ export const grokCodeTool = defineModelTool({
         const messages = [
             {
                 role: "system",
-                content: `You are Grok 4.3, expert programmer and code analyst.
+                content: `You are Grok 4.5, expert programmer and code analyst.
 Task: ${taskPrompts[task]}
 ${language ? `Language: ${language}` : ''}
 ${requirements ? `Requirements: ${requirements}` : ''}

package/dist/src/tools/openai-tools.js

@@ -143,8 +143,13 @@ reasoningEffort = "low", requireConfirmation = false, skipValidation = false) {
         }
         return { ...msg, content: validation.sanitized };
     });
-    // Model fallback chain - no fallbacks to test actual availability
+    // Model fallback chain
     const modelFallbacks = {
+        // gpt-5.6-sol is org-verification gated (403 "insufficient permissions" until verified);
+        // terra matches gpt-5.5 quality at half price, so it's the safe runtime fallback.
+        "gpt-5.6-sol": ["gpt-5.6-terra", "gpt-5.5"],
+        "gpt-5.6-terra": ["gpt-5.5"],
+        "gpt-5.6-luna": ["gpt-5.6-terra"],
         "gpt-5.4": [], // No fallback - test actual gpt-5.4
         "gpt-5.4-mini": ["gpt-5.4"], // Mini falls back to flagship
         "gpt-5.4-pro": [] // No fallback - test actual gpt-5.4-pro
@@ -204,8 +209,9 @@ reasoningEffort = "low", requireConfirmation = false, skipValidation = false) {
                 const error = await response.text();
                 lastError = `${currentModel}: ${response.statusText} - ${error}`;
                 console.error(`🔍 TRACE: ${currentModel} failed - Status: ${response.status}, Error: ${error}`);
-                // Check if it's a model not found error
-                if (response.status === 404 || error.includes('model') || error.includes('not found')) {
+                // Check if it's a model not found / not-yet-unlocked error
+                // (403 "insufficient permissions" = org-verification-gated tier, e.g. gpt-5.6-sol)
+                if (response.status === 404 || response.status === 403 || error.includes('model') || error.includes('not found') || error.includes('insufficient permissions')) {
                     console.error(`🔍 TRACE: Model ${currentModel} not available, trying fallback...`);
                     continue; // Try next model
                 }

package/dist/src/tools/provider-catalog.js

@@ -33,6 +33,7 @@ export const PROVIDER_GROUPS = [
             "grok_architect",
             "grok_brainstorm",
             "grok_search",
+            "grok_search_lite",
         ],
     },
     {

package/dist/src/tools/registry.js

@@ -39,6 +39,7 @@
 import { getAllPerplexityTools, isPerplexityAvailable } from "./perplexity-tools.js";
 import { getAllGrokTools, isGrokAvailable } from "./grok-tools.js";
 import { debugTriageTool } from "./debug-triage-tool.js";
+import { grokSearchLiteTool } from "./grok-search-lite-tool.js";
 import { isOpenAIAvailable, getAllOpenAITools } from "./openai-tools.js";
 import { specWriterTool } from "./spec-writer-tool.js";
 import { refinePromptTool } from "./refine-prompt-tool.js";
@@ -74,8 +75,10 @@ export async function getAllTools(inlineTools = []) {
     // 2) Grok (custom API).
     if (isGrokAvailable()) {
         tools.push(...getAllGrokTools());
-        // debug_triage — ranked-hypothesis bug triage (Grok 4.3) — gated on Grok.
+        // debug_triage — ranked-hypothesis bug triage (Grok flagship) — gated on Grok.
         tools.push(debugTriageTool);
+        // grok_search_lite — cheap live search (grok-4-1-fast, ~10x cheaper) — gated on Grok.
+        tools.push(grokSearchLiteTool);
     }
     // 3) OpenAI (GPT-5 suite). NB: getAllOpenAITools() ALSO self-guards (returns
     //    [] when unavailable); the outer guard is kept to mirror server.ts 1:1.

package/docs/TOOLS_REFERENCE.md

@@ -1,6 +1,6 @@
 # TachiBot MCP - Complete Tools Reference
 
-**Complete parameter schemas and example calls for all 64 tools** (65 registered — `continue_focus` is an always-on companion to `focus` outside the profile system; see [Meta & Orchestration](#meta--orchestration)).
+**Complete parameter schemas and example calls for all 65 tools** (66 registered — `continue_focus` is an always-on companion to `focus` outside the profile system; see [Meta & Orchestration](#meta--orchestration)).
 
 Schemas below are generated from the wire contract (`test/golden/__snapshots__/tool-contracts.json`) — the exact JSON Schema the MCP server publishes for each tool.
 
@@ -8,7 +8,7 @@ Schemas below are generated from the wire contract (`test/golden/__snapshots__/t
 
 ## Table of Contents
 
-- [Research & Search](#research--search) (5): [perplexity_ask](#perplexity_ask) · [perplexity_reason](#perplexity_reason) · [grok_search](#grok_search) · [openai_search](#openai_search) · [gemini_search](#gemini_search)
+- [Research & Search](#research--search) (6): [perplexity_ask](#perplexity_ask) · [perplexity_reason](#perplexity_reason) · [grok_search](#grok_search) · [grok_search_lite](#grok_search_lite) · [openai_search](#openai_search) · [gemini_search](#gemini_search)
 - [Reasoning & Planning](#reasoning--planning) (14): [grok_reason](#grok_reason) · [openai_reason](#openai_reason) · [qwen_reason](#qwen_reason) · [qwq_reason](#qwq_reason) · [kimi_thinking](#kimi_thinking) · [kimi_decompose](#kimi_decompose) · [deepseek_reason](#deepseek_reason) · [glm_reason](#glm_reason) · [stepfun_reason](#stepfun_reason) · [ernie_reason](#ernie_reason) · [planner_maker](#planner_maker) · [planner_runner](#planner_runner) · [list_plans](#list_plans) · [spec_writer](#spec_writer)
 - [Code Intelligence](#code-intelligence) (11): [kimi_code](#kimi_code) · [grok_code](#grok_code) · [grok_debug](#grok_debug) · [qwen_coder](#qwen_coder) · [qwen_algo](#qwen_algo) · [qwen_competitive](#qwen_competitive) · [deepseek_algo](#deepseek_algo) · [minimax_code](#minimax_code) · [minimax_agent](#minimax_agent) · [testgen](#testgen) · [debug_triage](#debug_triage)
 - [Analysis & Judgment](#analysis--judgment) (14): [gemini_analyze_text](#gemini_analyze_text) · [gemini_analyze_code](#gemini_analyze_code) · [gemini_judge](#gemini_judge) · [jury](#jury) · [diff_review](#diff_review) · [plan_critique](#plan_critique) · [gemini_brainstorm](#gemini_brainstorm) · [openai_brainstorm](#openai_brainstorm) · [openai_code_review](#openai_code_review) · [openai_explain](#openai_explain) · [grok_brainstorm](#grok_brainstorm) · [grok_architect](#grok_architect) · [security_review](#security_review) · [kimi_long_context](#kimi_long_context)
@@ -95,6 +95,30 @@ grok_search({
 
 ---
 
+### grok_search_lite
+
+Cheap web search (grok-4-1-fast, ~10x cheaper than grok_search). Same live-search pipeline and parameters as `grok_search` — use it for high-volume lookups and quick fact checks; use `grok_search` when synthesis quality matters most.
+
+#### Parameters
+
+| Parameter | Type | Required | Default | Description |
+|-----------|------|----------|---------|-------------|
+| `query` | `string` | ✅ Yes | - | Search query |
+| `max_search_results` | `number` | No | - | Max sources searched (costs per 1k) |
+| `recency` | `"all" \| "day" \| "week" \| "month" \| "year"` | No | - | Time filter |
+| `sources` | `Array<{type: "web"\|"news"\|"x"\|"rss", allowed_websites?: string[], country?: string}>` | No | - | Source configuration |
+
+#### Example
+
+```typescript
+grok_search_lite({
+  query: "current stable Node.js LTS version",
+  recency: "month"
+})
+```
+
+---
+
 ### openai_search
 
 Web search using GPT-5.4 with real-time web access. Put your QUERY in the `query` parameter.

package/docs/TOOL_PROFILES.md

@@ -25,11 +25,11 @@ Counts below are generated from `src/profiles/*.ts` via `npm run build:profiles`
 | Profile | Tools | Description |
 |---------|-------|-------------|
 | `minimal` | 13 | Minimal essential tools for basic tasks |
-| `research_power` | 35 | Research-focused with Grok search + all Perplexity + brainstorming |
+| `research_power` | 36 | Research-focused with Grok search + all Perplexity + brainstorming |
 | `code_focus` | 42 | Code-heavy work with debugging and analysis |
-| `balanced` | 53 | Balanced set for general use |
-| `heavy_coding` | 57 | Heavy coding with all reasoning & code tools |
-| `full` | 64 | **Default profile** — every tool enabled |
+| `balanced` | 54 | Balanced set for general use |
+| `heavy_coding` | 58 | Heavy coding with all reasoning & code tools |
+| `full` | 65 | **Default profile** — every tool enabled |
 
 ---
 
@@ -50,13 +50,13 @@ Minimal essential tools for basic tasks.
 
 ---
 
-### 2. `research_power` (35 tools)
+### 2. `research_power` (36 tools)
 
 Research-focused with Grok search + all Perplexity + brainstorming + the full reasoning/planning suite.
 
 **Tools:**
 - Meta: `think`, `focus`, `tachi`, `doctor`, `nextThought`, `usage_stats`
-- Research: `perplexity_ask`, `perplexity_reason`, `grok_search`, `openai_search`, `gemini_search`
+- Research: `perplexity_ask`, `perplexity_reason`, `grok_search`, `grok_search_lite`, `openai_search`, `gemini_search`
 - Reasoning: `grok_reason`, `qwq_reason`, `qwen_reason`, `kimi_thinking`, `kimi_decompose`, `deepseek_reason`, `glm_reason`, `stepfun_reason`, `ernie_reason`
 - Judgment: `gemini_judge`, `jury`
 - Creative: `openai_brainstorm`, `gemini_brainstorm`
@@ -68,7 +68,7 @@ Research-focused with Grok search + all Perplexity + brainstorming + the full re
 **Best for:** Deep research, fact-checking, multi-source verification, brainstorming sessions.
 
 **Key tools for research:**
-- `grok_search` / `openai_search` / `gemini_search` — live web search, three different engines
+- `grok_search` / `openai_search` / `gemini_search` — live web search, three different engines (`grok_search_lite` = same Grok search on grok-4-1-fast, ~10x cheaper for high-volume lookups)
 - `jury` — multi-model consensus (parallel jurors + Gemini synthesis)
 - `gemini_judge` — evaluate/synthesize/rank/resolve multiple perspectives into one verdict
 
@@ -94,13 +94,13 @@ Code-heavy work with debugging and analysis.
 
 ---
 
-### 4. `balanced` (53 tools)
+### 4. `balanced` (54 tools)
 
 Balanced set for general use — nearly everything except the deepest niche reasoning/creative extras.
 
 **Tools:**
 - Meta: `think`, `focus`, `tachi`, `doctor`, `nextThought`, `usage_stats`
-- Research: `perplexity_ask`, `perplexity_reason`, `grok_search`, `openai_search`, `gemini_search`
+- Research: `perplexity_ask`, `perplexity_reason`, `grok_search`, `grok_search_lite`, `openai_search`, `gemini_search`
 - Reasoning: `grok_reason`, `qwq_reason`, `qwen_reason`, `kimi_thinking`, `kimi_decompose`, `deepseek_reason`, `glm_reason`, `stepfun_reason`, `ernie_reason`
 - Code: `grok_code`, `qwen_coder`, `qwen_algo`, `kimi_code`, `kimi_long_context`, `minimax_code`, `minimax_agent`, `deepseek_algo`, `debug_triage`
 - Judgment: `gemini_judge`, `jury`
@@ -117,13 +117,13 @@ Balanced set for general use — nearly everything except the deepest niche reas
 
 ---
 
-### 5. `heavy_coding` (57 tools)
+### 5. `heavy_coding` (58 tools)
 
 Heavy coding with the full reasoning & code toolset, minus the general workflow-management tools.
 
 **Tools:**
 - Meta: `think`, `focus`, `tachi`, `doctor`, `nextThought`, `usage_stats`
-- Research: `perplexity_ask`, `perplexity_reason`, `grok_search`, `openai_search`, `gemini_search`
+- Research: `perplexity_ask`, `perplexity_reason`, `grok_search`, `grok_search_lite`, `openai_search`, `gemini_search`
 - Reasoning: `grok_reason`, `openai_reason`, `qwq_reason`, `qwen_reason`, `kimi_thinking`, `kimi_decompose`, `deepseek_reason`, `glm_reason`, `stepfun_reason`
 - Code: `grok_code`, `grok_debug`, `grok_architect`, `openai_code_review`, `openai_explain`, `gemini_analyze_code`, `qwen_coder`, `qwen_algo`, `kimi_code`, `kimi_long_context`, `minimax_code`, `minimax_agent`, `deepseek_algo`, `debug_triage`
 - Judgment: `gemini_judge`, `jury`
@@ -141,7 +141,7 @@ Heavy coding with the full reasoning & code toolset, minus the general workflow-
 
 ---
 
-### 6. `full` (64 tools) — default
+### 6. `full` (65 tools) — default
 
 All tools enabled for maximum capability. This is the profile set in `tools.config.json`'s `activeProfile` by default.
 
@@ -149,7 +149,7 @@ All tools enabled for maximum capability. This is the profile set in `tools.conf
 
 **Best for:** Maximum flexibility, specialized tasks, demonstrations.
 
-**Note:** tools still self-gate on API keys — with no `OPENROUTER_API_KEY`, the Qwen/Kimi/MiniMax/DeepSeek/GLM/StepFun/ERNIE tools won't register even under `full`. Run `doctor` to see exactly what's active. `continue_focus` is always registered regardless of profile (see [TOOLS_REFERENCE.md](TOOLS_REFERENCE.md#continue_focus)) and isn't counted in the 64.
+**Note:** tools still self-gate on API keys — with no `OPENROUTER_API_KEY`, the Qwen/Kimi/MiniMax/DeepSeek/GLM/StepFun/ERNIE tools won't register even under `full`. Run `doctor` to see exactly what's active. `continue_focus` is always registered regardless of profile (see [TOOLS_REFERENCE.md](TOOLS_REFERENCE.md#continue_focus)) and isn't counted in the 65.
 
 ---
 

package/package.json

@@ -2,7 +2,7 @@
   "name": "tachibot-mcp",
   "mcpName": "io.github.byPawel/tachibot-mcp",
   "displayName": "TachiBot MCP - Universal AI Orchestrator",
-  "version": "2.26.1",
+  "version": "2.27.0",
   "type": "module",
   "main": "dist/src/server.js",
   "bin": {
@@ -76,7 +76,7 @@
     "node": ">=22.0.0"
   },
   "license": "AGPL-3.0",
-  "description": "TachiBot MCP: 51 AI tools, 7 providers. Multi-model orchestration (Perplexity, Grok, OpenAI, Gemini, Qwen, Kimi, MiniMax), YAML workflows, token-optimized profiles. Smart routing, parallel execution, jury system.",
+  "description": "TachiBot MCP: 65 AI tools, 12 providers. Multi-model orchestration (Perplexity, Grok, OpenAI, Gemini, Qwen, Kimi, MiniMax, DeepSeek, GLM, StepFun, ERNIE, local), YAML workflows, token-optimized profiles. Smart routing, parallel execution, jury system.",
   "dependencies": {
     "@types/node": "^22.13.17",
     "@types/react": "^19.2.7",

package/profiles/balanced.json

@@ -1,5 +1,5 @@
 {
-  "description": "Balanced set for general use (53 tools)",
+  "description": "Balanced set for general use (54 tools)",
   "tools": {
     "think": true,
     "focus": true,
@@ -16,6 +16,7 @@
     "grok_architect": false,
     "grok_brainstorm": false,
     "grok_search": true,
+    "grok_search_lite": true,
     "openai_reason": true,
     "openai_brainstorm": true,
     "openai_code_review": false,

package/profiles/code_focus.json

@@ -16,6 +16,7 @@
     "grok_architect": false,
     "grok_brainstorm": false,
     "grok_search": false,
+    "grok_search_lite": false,
     "openai_reason": false,
     "openai_brainstorm": false,
     "openai_code_review": true,

package/profiles/full.json

@@ -1,5 +1,5 @@
 {
-  "description": "Default profile — all tools enabled for maximum capability (64 tools)",
+  "description": "Default profile — all tools enabled for maximum capability (65 tools)",
   "tools": {
     "think": true,
     "focus": true,
@@ -16,6 +16,7 @@
     "grok_architect": true,
     "grok_brainstorm": true,
     "grok_search": true,
+    "grok_search_lite": true,
     "openai_reason": true,
     "openai_brainstorm": true,
     "openai_code_review": true,

package/profiles/heavy_coding.json

@@ -1,5 +1,5 @@
 {
-  "description": "Heavy coding with all reasoning & code tools (57 tools)",
+  "description": "Heavy coding with all reasoning & code tools (58 tools)",
   "tools": {
     "think": true,
     "focus": true,
@@ -16,6 +16,7 @@
     "grok_architect": true,
     "grok_brainstorm": true,
     "grok_search": true,
+    "grok_search_lite": true,
     "openai_reason": true,
     "openai_brainstorm": true,
     "openai_code_review": true,

package/profiles/minimal.json

@@ -16,6 +16,7 @@
     "grok_architect": false,
     "grok_brainstorm": false,
     "grok_search": false,
+    "grok_search_lite": false,
     "openai_reason": false,
     "openai_brainstorm": false,
     "openai_code_review": false,

package/profiles/research_power.json

@@ -1,5 +1,5 @@
 {
-  "description": "Research-focused with Grok search + all Perplexity + brainstorming (35 tools)",
+  "description": "Research-focused with Grok search + all Perplexity + brainstorming (36 tools)",
   "tools": {
     "think": true,
     "focus": true,
@@ -16,6 +16,7 @@
     "grok_architect": false,
     "grok_brainstorm": false,
     "grok_search": true,
+    "grok_search_lite": true,
     "openai_reason": false,
     "openai_brainstorm": true,
     "openai_code_review": false,

package/tools.config.json

@@ -23,7 +23,8 @@
       "debug_triage",
       "grok_architect",
       "grok_brainstorm",
-      "grok_search"
+      "grok_search",
+      "grok_search_lite"
     ],
     "openai": ["openai_brainstorm", "openai_reason", "openai_code_review", "openai_explain", "openai_search", "spec_writer", "refine_prompt"],
     "gemini": [
@@ -73,6 +74,7 @@
       "grok_architect": true,
       "grok_brainstorm": true,
       "grok_search": true,
+      "grok_search_lite": true,
       "openai_brainstorm": true,
       "openai_reason": true,
       "openai_code_review": true,