Signed-off-by: Matt Bruce <mbrucedogs@gmail.com>

openclaw-setup-max/PRD.md

@@ -154,7 +154,7 @@ Outputs:
 ## 10) Acceptance Criteria
 
 - AC-1: Running `bash ./scripts/model_profile_switch.sh free` changes default to `ollama/qwen3:14b`.
-- AC-2: Running `bash ./scripts/model_profile_switch.sh paid` restores default to `moonshot/kimi-k2.5`.
+- AC-2: Running `bash ./scripts/model_profile_switch.sh paid` restores default to `openrouter/moonshotai/kimi-k2.5`.
 - AC-3: With schedule enabled (`21`/`7`), profile changes correctly by local time window.
 - AC-4: If operator manually drifts model during scheduled window, schedule guard re-aligns on next run.
 - AC-5: Budget guard warns and reverts when high model remains active beyond thresholds.
@@ -186,4 +186,3 @@ Outputs:
 - Per-channel session key mapping (Telegram/Discord-specific switching).
 - Daily spend estimator from model usage telemetry.
 - Optional provider locking profile policies.
-

openclaw-setup-max/README.md

@@ -1,7 +1,7 @@
 # OpenClaw Setup Max (Paid + Free Model Switching)
 
 This workspace runs OpenClaw with:
-- A paid high-quality model profile (`moonshot/kimi-k2.5`)
+- A paid high-quality model profile (`openrouter/moonshotai/kimi-k2.5`)
 - A free local profile (Ollama-only)
 - One-command live switching
 - Optional automatic day/night switching (example: free from 9pm-7am)
@@ -27,7 +27,7 @@ Everything else is grouped by function:
 ## 1) What You Get
 
 - `paid` profile:
-  - Primary: `moonshot/kimi-k2.5`
+  - Primary: `openrouter/moonshotai/kimi-k2.5`
   - Fallbacks: `ollama/qwen3:14b`, `ollama/llama3.2:3b`
 - `free` profile:
   - Primary: `ollama/qwen3:14b`
@@ -102,6 +102,17 @@ Check current model state:
 bash ./scripts/model_profile_switch.sh status
 ```
 
+Run a web research task from terminal:
+
+```bash
+./scripts/research.sh latest swiftui observation patterns
+```
+
+Optional flags:
+- `--sources <count>` minimum source count (default `3`)
+- `--thinking <level>` sets OpenClaw thinking level
+- `--deliver` sends the final response back to the current channel
+
 ## 5) How Live Switching Works
 
 When you run:

openclaw-setup-max/config/model-budget-guard.config.json

@@ -3,8 +3,9 @@
   "sessionKey": "agent:main:main",
   "lowModel": "ollama/qwen3:14b",
   "highModels": [
-    "moonshot/kimi-k2.5",
-    "xai/grok-4-fast"
+    "openrouter/moonshotai/kimi-k2.5",
+    "xai/grok-4-fast",
+    "xai/grok-code-fast-1"
   ],
   "warnAfterMinutes": 2,
   "revertAfterMinutes": 45,

openclaw-setup-max/config/model-profiles.config.json

@@ -2,7 +2,7 @@
   "profiles": {
     "paid": {
       "description": "High-quality paid model for deep work",
-      "primary": "moonshot/kimi-k2.5",
+      "primary": "openrouter/moonshotai/kimi-k2.5",
       "fallbacks": [
         "ollama/qwen3:14b",
         "ollama/llama3.2:3b"

openclaw-setup-max/docs/operations/MODEL_SWITCHING.md

@@ -2,7 +2,7 @@
 
 This setup adds two model profiles:
 
-- `paid` -> `moonshot/kimi-k2.5` with local Ollama fallbacks
+- `paid` -> `openrouter/moonshotai/kimi-k2.5` with local Ollama fallbacks
 - `free` -> local Ollama-only stack
 
 ## On-demand switch (immediate)

openclaw-setup-max/scripts/research.sh

@@ -0,0 +1,108 @@
+#!/usr/bin/env bash
+set -euo pipefail
+
+usage() {
+  cat <<'USAGE'
+Usage:
+  research <topic...>
+  research --sources <count> <topic...>
+  research --thinking <off|minimal|low|medium|high> <topic...>
+  research --deliver <topic...>
+
+Examples:
+  research swiftui state management in 2026
+  research --sources 5 best local llm coding setups
+  research --thinking high ai coding agents comparison
+USAGE
+}
+
+source_count="3"
+thinking_level=""
+deliver="false"
+agent_id="${RESEARCH_AGENT_ID:-main}"
+
+while (( "$#" )); do
+  case "$1" in
+    --sources)
+      shift
+      if [[ $# -eq 0 ]]; then
+        echo "[research] --sources requires a number" >&2
+        usage
+        exit 1
+      fi
+      source_count="$1"
+      ;;
+    --thinking)
+      shift
+      if [[ $# -eq 0 ]]; then
+        echo "[research] --thinking requires a level" >&2
+        usage
+        exit 1
+      fi
+      thinking_level="$1"
+      ;;
+    --deliver)
+      deliver="true"
+      ;;
+    -h|--help)
+      usage
+      exit 0
+      ;;
+    --)
+      shift
+      break
+      ;;
+    -*)
+      echo "[research] unknown argument: $1" >&2
+      usage
+      exit 1
+      ;;
+    *)
+      break
+      ;;
+  esac
+  shift
+done
+
+if [[ $# -eq 0 ]]; then
+  echo "[research] topic is required" >&2
+  usage
+  exit 1
+fi
+
+if ! [[ "$source_count" =~ ^[0-9]+$ ]] || [[ "$source_count" -lt 1 ]]; then
+  echo "[research] --sources must be a positive integer" >&2
+  exit 1
+fi
+
+if ! command -v openclaw >/dev/null 2>&1; then
+  echo "[research] openclaw CLI is required" >&2
+  exit 1
+fi
+
+topic="$*"
+message=$(
+  cat <<EOF
+Research topic: $topic
+
+Instructions:
+- Search the web for up-to-date, reputable sources.
+- Review at least $source_count distinct sources.
+- Summarize the key points in concise bullet points.
+- Include a short "Sources" section with direct links.
+- Mention relevant publish/event dates when time-sensitive.
+EOF
+)
+
+cmd=(openclaw agent --channel last --message "$message")
+if [[ -n "$agent_id" ]]; then
+  cmd+=(--agent "$agent_id")
+fi
+if [[ -n "$thinking_level" ]]; then
+  cmd+=(--thinking "$thinking_level")
+fi
+if [[ "$deliver" == "true" ]]; then
+  cmd+=(--deliver)
+fi
+
+"${cmd[@]}"