Master Ops

---
name: solcex-ops-master
description: >
  System Operations Manual for Akash container services at SolCex Exchange.
  Technical reference documentation for all running services, APIs, intelligence sources,
  cron scheduling, pipeline management, and coordination protocols.
  v6.0.19 — Indonesia Sprint Day 8 Edition (Mar 2, 2026).
  OpenClaw 2026.2.26 | 5 Parallel Sub-Agents LIVE | Orchestrator wired to 4 scan crons
  Twitter Bot v3.0 Sales Funnel | MiniMax M2.5 primary + bankr/gpt-5-nano sub-agents
  Bankr LLM Gateway (8 models, dual API keys) | 40 cron jobs | 16/17 intelligence sources
  AgentProof integration in progress (Avalanche registration pending)
  Reporting rule: every completed task reports to Telegram.
---

# SYSTEM OPERATIONS MANUAL v6.0.19 — Indonesia Sprint Day 8

> **Technical reference documentation** for all services running in this Akash container.
> Covers team contacts, service configuration, APIs, intelligence sources, sub-agent architecture,
> cron scheduling, pipeline management, scoring, revenue tracking, and operational procedures.
> Updated Mar 2, 2026 from Jakarta, Indonesia.

---

## 1. TEAM CONTACTS

| Role                | Identity                                                   | Responsibility                                           |
| ------------------- | ---------------------------------------------------------- | -------------------------------------------------------- |
| **Ogie**            | Operations Lead, Inflight Chef (Saudia) + BD Lead (SolCex) | Strategy, approvals, manual Twitter, partnerships        |
| **Claude Opus 4.6** | Strategy/ops advisor + dev partner                         | Documentation, analysis, planning, code, outreach drafts |
| **Alexander**       | SolCex team member                                         | Support (NOT primary decision-maker)                     |

**Ogie Comms:** Telegram @Ogie2 (Chat ID: 950395553) | Twitter: @hidayahanka1
**Service Twitter:** @BuzzBySolCex
**Other accounts:** @SolCex_Exchange, @SolCex_Academy, @Solcex_intern, @arloxshot, @Alexanderbtcc, @Dinozzolo
**LinkedIn:** https://www.linkedin.com/in/howtobecomeachef/

---

## 2. SOLCEX LISTING PACKAGE (Reference Data)

| Component             | Value                                               |
| --------------------- | --------------------------------------------------- |
| Total listing package | 15K USDT (5K fee + 10K liquidity)                   |
| Market making         | Included (3-month, $450K+ depth, 0.15% spread)      |
| Whale airdrop         | Included (450+ whale traders, avg portfolio $500K+) |
| Listing timeline      | Fast-track 10-14 days                               |
| Ogie's commission     | $1K per listing (**NEVER share publicly**)          |

**Public-facing listing info (for Twitter LIST replies):** Show benefits only (market making, whale airdrop, fast-track, dedicated support). Push pricing discussion to DM with Ogie.

---

## 3. CONTAINER SERVICES & DEPLOYMENT

### 3.1 ERC-8004 On-Chain Identity

| Chain           | Agent ID | Registry                                       |
| --------------- | -------- | ---------------------------------------------- |
| **Ethereum**    | #25045   | 0x8004A818BFB912233c491871b3d84c89A494BD9e     |
| **Base**        | #17483   | 0x8004A169FB4a3325136EB29fA0ceB6D2e539a432     |
| **Base (anet)** | #18709   | Registered Feb 22, 2026 — 4 skills             |
| **Avalanche**   | Pending  | AgentProof registration (0.1 AVAX bond funded) |
| **Colosseum**   | #3734    | Agent Hackathon entry                          |

**Reputation Registries:**

- Ethereum: 0x8004B663056A597Dffe9eCcC1965A193B7388713
- Base: 0x8004BAa17C55a88189AE136b182e5fdA19dE9b63

### 3.2 Deployment Stack

| Field                  | Value                                                                         |
| ---------------------- | ----------------------------------------------------------------------------- |
| **Platform**           | Akash Network (decentralized cloud)                                           |
| **Container**          | Docker via Akash SDL, tini entrypoint                                         |
| **Docker Image**       | `ghcr.io/buzzbysolcex/buzz-bd-agent:v6.0.19`                                  |
| **Container Registry** | GitHub Container Registry (ghcr.io) — PUBLIC                                  |
| **Runtime**            | **OpenClaw v2026.2.26**                                                       |
| **Resources**          | 2 CPU, 4GB RAM, 10GB persistent storage                                       |
| **Cost**               | ~$4-8/month                                                                   |
| **Process**            | `openclaw gateway --port 18789 --allow-unconfigured` (foreground, tini PID 1) |
| **State dir**          | `/data/.openclaw/` (persistent across restarts)                               |
| **Workspace**          | `/data/workspace/` (persistent)                                               |
| **Config**             | `/data/.openclaw/openclaw.json` (auto-generated from env vars on boot)        |
| **Skills dir**         | `/data/workspace/skills/` + `/root/.openclaw/workspace/skills/`               |
| **Telegram bot**       | @BuzzBySolCex_bot (dmPolicy: open, allowFrom: ["*"])                          |
| **Telegram commands**  | 48 registered via Bot API                                                     |
| **GitHub (agent)**     | github.com/buzzbysolcex/buzz-bd-agent                                         |
| **GitHub (plugin)**    | github.com/buzzbysolcex/plugin-solcex-bd                                      |

### 3.3 Wallets

| Name              | Chain      | Address                                        | Purpose                                             |
| ----------------- | ---------- | ---------------------------------------------- | --------------------------------------------------- |
| **anet**          | Base (EVM) | `0x2Dc03124091104E7798C0273D96FC5ED65F05aA9`   | Main wallet, Bankr fee recipient, Bankr LLM credits |
| **Deploy wallet** | Base (EVM) | `0xfa04c7d627ba707a1ad17e72e094b45150665593`   | Bankr token deployment                              |
| **Bankr Trading** | Base (EVM) | `0x8ea088f4a206d9373a4f6dffb12e2c7e5583b967`   | Bankr trading wallet (auto-provisioned)             |
| **Buzz Base**     | Base (EVM) | `0x4b362B7db6904A72180A37307191fdDc4eD282Ab`   | Base operations                                     |
| **Lobster**       | Solana     | `5iC7pGyzqpXD2xTK4Ww7zKRDVo9cceyHNeKBTiemo5Jp` | Solana operations                                   |
| **AgentProof**    | Avalanche  | MetaMask (0.657 AVAX funded)                   | AgentProof registration                             |

**Deprecated wallets (DO NOT USE):**

- ClawRouter `0x56f76...5C980` — REMOVED v6.0.17
- BlockRun x402 `0x6ea36...5493b` — Key lost, wallet empty, REMOVED v6.0.17
- ClawRouter BlockRun `0x9b289...3A76` — Key lost, REMOVED

### 3.4 OpenClaw Version History

| Version        | Date         | Key Change                                                                                             |
| -------------- | ------------ | ------------------------------------------------------------------------------------------------------ |
| **v2026.2.26** | Feb 28, 2026 | **CURRENT** — MiniMax auth header fix, Telegram fixes, cron reliability, sessions_spawn for sub-agents |
| v2026.2.24     | Feb 25, 2026 | Model fallback chain fix, OpenRouter auth fix, Telegram IPv4, cron delivery fixes                      |

### 3.5 MiniMax M2.5 Configuration

**MiniMax M2.5 MUST use `anthropic-messages` API format.**

| Setting           | Value                              |
| ----------------- | ---------------------------------- |
| **baseUrl**       | `https://api.minimax.io/anthropic` |
| **api**           | `anthropic-messages`               |
| **maxTokens**     | `8192`                             |
| **contextWindow** | `200000`                           |

### 3.6 Docker Image & Pipeline

| Field           | Value                                         |
| --------------- | --------------------------------------------- |
| **Base image**  | `node:22-slim`                                |
| **OpenClaw**    | `npm install -g openclaw@2026.2.26`           |
| **Bankr CLI**   | `npm install -g @bankr/cli`                   |
| **Current tag** | `v6.0.19`                                     |
| **Registry**    | `ghcr.io/buzzbysolcex/buzz-bd-agent` (PUBLIC) |

**Update procedure:**

```bash
cd ~/buzz-bd-agent
docker builder prune -f
docker build --no-cache -t ghcr.io/buzzbysolcex/buzz-bd-agent:vNEW .
docker push ghcr.io/buzzbysolcex/buzz-bd-agent:vNEW
# Akash Console: Close old → Create New Deployment with updated SDL
# ALWAYS use unique image tags — Akash providers cache by tag name
```

### 3.7 Telegram Commands (48 Registered)

**System (8):** help, status, health, version, config, memory, crons, experience
**Scanning (9):** scan, score, search, verify, trending, boosts, aixbt, rugcheck, socials
**Pipeline (7):** pipeline, report, weekly, digest, alpha, listings, competitors
**Wallet/Intel (6):** wallet, forensics, whales, dflow, budget, smartmoney
**Outreach (7):** outreach, approve, reject, warmup, followup, breakup, dealsheet, postlisting
**Bankr Deploy (3):** launch, deploys, deploy_stats
**Twitter/X (3):** tweet, thread, engage
**Personal (2):** prayer, sprint
**Emergency (5):** sos, stop, stop_email, stop_scan, resume

---

## 4. LLM CASCADE (Cost Control)

### 4.1 Provider Configuration

| Priority               | Model               | Provider          | API Format         | Cost               | Role                      |
| ---------------------- | ------------------- | ----------------- | ------------------ | ------------------ | ------------------------- |
| **PRIMARY**            | MiniMax M2.5 (229B) | MiniMax Direct    | anthropic-messages | ~$41/mo            | Orchestrator + main agent |
| **SUB-AGENT DEFAULT**  | GPT-5 Nano          | Bankr LLM Gateway | openai-completions | $0.10/$0.40 per 1M | All 5 sub-agents          |
| **SUB-AGENT FALLBACK** | Claude Haiku 4.5    | Bankr LLM Gateway | anthropic-messages | $0.80/$4.00 per 1M | Sub-agent fallback        |
| **MAIN FALLBACK 1**    | Gemini 3 Flash      | Bankr LLM Gateway | openai-completions | $0.15/$0.60 per 1M | Main agent fallback       |
| **MAIN FALLBACK 2**    | Claude Haiku 4.5    | Bankr LLM Gateway | anthropic-messages | $0.80/$4.00 per 1M | Main agent fallback       |
| **MAIN FALLBACK 3**    | GPT-5 Nano          | Bankr LLM Gateway | openai-completions | $0.10/$0.40 per 1M | Main agent fallback       |

**IMPORTANT:** gemini-3-flash has a 0 output token bug in sub-agent context. DO NOT use for sub-agents. Use gpt-5-nano (confirmed working) with claude-haiku-4.5 fallback.

**Main agent cascade:** MiniMax → Gemini 3 Flash → Haiku 4.5 → GPT-5 Nano
**Sub-agent cascade:** GPT-5 Nano → Claude Haiku 4.5 (NO gemini-3-flash)

### 4.2 Bankr LLM Gateway (Dual API Keys — v6.0.19)

| Field                | Value                                                 |
| -------------------- | ----------------------------------------------------- |
| **Gateway URL**      | `https://llm.bankr.bot`                               |
| **Key 1 (Partner)**  | `bk_JSCN...` — Partner API (token deploys, agent API) |
| **Key 2 (LLM)**      | `bk_M94Y...` — LLM Gateway (sub-agent inference)      |
| **API Format**       | OpenAI-compatible (openai-completions)                |
| **Credits**          | $15 initial balance (funded Mar 1, 2026)              |
| **Dashboard**        | bankr.bot/llm                                         |
| **Auto Top-Up**      | Available (enable when credits < $1)                  |
| **Models Available** | 8 total                                               |

**All Bankr LLM Models:**

| Model             | Context | Input/Output Cost (per 1M) | Best For                                       |
| ----------------- | ------- | -------------------------- | ---------------------------------------------- |
| gpt-5-nano        | 400K    | $0.10 / $0.40              | **Sub-agent default** (cheapest working)       |
| gemini-3-flash    | 1M      | $0.15 / $0.60              | Main fallback ONLY (⚠️ 0 output in sub-agents) |
| claude-haiku-4.5  | 200K    | $0.80 / $4.00              | Sub-agent fallback + quality tasks             |
| gpt-5-mini        | 400K    | $0.40 / $1.60              | General purpose                                |
| qwen3-coder       | 262K    | $0.30 / $1.20              | Code generation                                |
| kimi-k2.5         | 262K    | $0.60 / $2.40              | Long context analysis                          |
| gemini-3-pro      | 1M      | $1.25 / $10.00             | Complex reasoning                              |
| claude-sonnet-4.6 | 200K    | $3.00 / $15.00             | Premium/complex tasks                          |

**Self-Sustaining Loop:**
Token deploy fees → fund LLM credits → power inference → discover more tokens → loop
Break-even: ~$5-10/mo in LLM credits = 1-2 successful token deploys

### 4.3 Removed Providers (DO NOT USE)

| Provider                   | Reason                        | Removed |
| -------------------------- | ----------------------------- | ------- |
| ClawRouter / BlockRun x402 | Replaced by Bankr LLM Gateway | v6.0.17 |
| OpenRouter                 | Dead API key                  | v6.0.17 |
| AkashML (Qwen)             | 401 Unauthorized              | v6.0.17 |

**Port 8402 (ClawRouter proxy) has been removed from the SDL.**

---

## 5. SUB-AGENT ARCHITECTURE (NEW v6.0.19)

### 5.1 Overview

5 parallel sub-agents running on Akash via OpenClaw `sessions_spawn`. Each sub-agent runs in an isolated session, executes its task, and announces results back to the Orchestrator (Buzz main agent).

**Built with Claude Code + Superpowers (obra/superpowers) — 602 tests passing.**

### 5.2 Sub-Agent Registry

| Label             | Class        | Layer        | Sources                                      | Weight      | Model            |
| ----------------- | ------------ | ------------ | -------------------------------------------- | ----------- | ---------------- |
| **scanner-agent** | ScannerAgent | L1 Discovery | DexScreener, GeckoTerminal, AIXBT, DS Boosts | Entry point | bankr/gpt-5-nano |
| **safety-agent**  | SafetyAgent  | L2 Filter    | RugCheck, DFlow MCP                          | 0.30        | bankr/gpt-5-nano |
| **wallet-agent**  | WalletAgent  | L2 Filter    | Helius, Allium                               | 0.30        | bankr/gpt-5-nano |
| **social-agent**  | SocialAgent  | L3 Research  | Grok, Serper, ATV, Firecrawl                 | 0.20        | bankr/gpt-5-nano |
| **scorer-agent**  | ScorerAgent  | L4 Score     | 100-point composite (11 factors)             | 0.20        | bankr/gpt-5-nano |

**Orchestrator** (Buzz main on MiniMax M2.5) dispatches all 5 in parallel.

**Weights:** safety(0.30) + wallet(0.30) + social(0.20) + scorer(0.20) = 1.0

**Note:** DeployAgent removed — Bankr Partner API handles deploys via Twitter Bot v3.0 DEPLOY route.

### 5.3 Sub-Agent Configuration

```json
{
  "agents": {
    "defaults": {
      "subagents": {
        "model": "bankr/gpt-5-nano",
        "runTimeoutSeconds": 60,
        "archiveAfterMinutes": 60
      }
    }
  }
}
```

**OpenClaw sub-agent limits:**

- maxChildrenPerAgent: 5 (we use exactly 5)
- maxSpawnDepth: 1 (sub-agents cannot spawn sub-agents)
- Auto-archive after 60 minutes
- Each sub-agent has own context and token usage

### 5.4 Orchestrator Flow

```
Cron fires (#1-4: scan-morning/midday/evening/night)
  │
  ▼
Orchestrator runs buzz-pipeline-scan (L1 Discovery)
  │ → DexScreener multi-query (new, pump, meme, ai, cat, dog)
  │ → Filter: MC $10K-$100M, Liq >$1K, exclude base assets
  │ → 80+ candidates → top 3-5 selected for parallel scoring
  │
  ▼
sessions_spawn → 5 sub-agents in parallel (gpt-5-nano)
  │
  ├── scanner-agent: Full token data (DexScreener pairs endpoint)
  ├── safety-agent: RugCheck + contract verification
  ├── wallet-agent: Helius deployer forensics + holder analysis
  ├── social-agent: Grok sentiment + Serper web search
  └── scorer-agent: 100-point composite from all agent data
  │
  ▼
All 5 announce results back (timeout: 60-120s)
  │
  ▼
Orchestrator aggregates weighted scores
  │ → safety(0.30) + wallet(0.30) + social(0.20) + scorer(0.20)
  │ → Same 100-point scale, same thresholds
  │
  ▼
Compiled report → Telegram
  │ → 🔥 HOT (85+) / ✅ QUAL (70+) / 👀 WATCH (50+) / ❌ SKIP (<50)
  │
  ▼
Results saved → /data/workspace/memory/pipeline/latest-scan.json
```

### 5.5 Skills (OpenClaw Workspace)

| Skill                  | Location                                                       | Function                    |
| ---------------------- | -------------------------------------------------------------- | --------------------------- |
| **buzz-pipeline-scan** | `/root/.openclaw/workspace/skills/buzz-pipeline-scan/scan.js`  | L1-L5 pipeline scan         |
| **orchestrator**       | `/root/.openclaw/workspace/skills/orchestrator/orchestrate.js` | Parallel sub-agent dispatch |
| atv-batch-skill.js     | workspace/skills/                                              | ATV identity verification   |
| bankr                  | workspace/skills/                                              | Bankr token deploy          |
| bankr-partner          | workspace/skills/                                              | Bankr partner API           |
| bankr-signals          | workspace/skills/                                              | Bankr trading signals       |
| buzz-scan-formatter    | workspace/skills/                                              | Scan output formatting      |
| clawrouter             | workspace/skills/                                              | (legacy)                    |
| content-filter         | workspace/skills/                                              | Content moderation          |
| data-failover          | workspace/skills/                                              | API failover handling       |
| master-ops             | workspace/skills/                                              | Master Ops reference        |
| quillshield            | workspace/skills/                                              | Contract safety             |
| twitter-poster         | workspace/skills/                                              | Twitter posting             |

### 5.6 First Orchestrated Scan Results (Mar 2, 2026)

```
🔍 ORCHESTRATED SCAN: $GORK — March 2, 2026

📡 SCANNER (L1): ✅ 5 boosted tokens found (24K tokens)
🛡️ SAFETY (L2): ✅ RugCheck fetched (14K tokens)
👛 WALLET (L2): ⚠️ Completed, 0 output (fallback issue)
🌐 SOCIAL (L3): ❌ Timed out (needs 120s)
📊 SCORER (L4): ✅ 48/100 — WATCH (6.8K tokens)

4/5 agents completed. Parallel execution confirmed.
```

---

## 6. API KEYS & CREDENTIALS

### 6.1 Currently Active

| #   | Service                       | Status                                     |
| --- | ----------------------------- | ------------------------------------------ |
| 1   | **MiniMax (Direct)**          | ✅ PRIMARY LLM (Orchestrator)              |
| 2   | **Bankr LLM Gateway (Key 2)** | ✅ SUB-AGENT LLM (gpt-5-nano, $15 credits) |
| 3   | **Telegram**                  | ✅ Bot token                               |
| 4   | **Bankr Partner (Key 1)**     | ✅ Token deploy (bk_JSCN...)               |
| 5   | **Helius**                    | ✅ Rate limited at volume                  |
| 6   | **Firecrawl**                 | ✅ FREE plan (500 credits)                 |
| 7   | **Grok/xAI**                  | ✅ Configured                              |
| 8   | **ATV Web3 Identity**         | ✅ FIXED — api.web3identity.com            |
| 9   | **Serper**                    | ✅ Configured                              |
| 10  | **X API v2**                  | ✅ Tweet posting + mentions                |
| 11  | **Allium**                    | ✅ Configured                              |
| 12  | **Hyperbrowser**              | ✅ Configured                              |
| 13  | **Gmail**                     | ✅ OAuth configured                        |
| 14  | **Moltbook**                  | ✅ Agent registered                        |
| 15  | **Google API**                | ✅ Configured                              |

### 6.2 Known Issues

| Service        | Issue                                                            |
| -------------- | ---------------------------------------------------------------- |
| Allium         | Endpoint returning 404                                           |
| Nansen x402    | x402-client not installed — returns NO_DATA                      |
| gemini-3-flash | 0 output tokens in sub-agent context — DO NOT use for sub-agents |
| RugCheck       | DNS intermittent — some scans fail to resolve                    |
| Prayer crons   | dmPolicy blocking delivery — gateway restart needed              |

### 6.3 Security Rules

- **NEVER** include keys in public repos/GitHub
- **NEVER** share in Telegram/Twitter/Moltbook
- Firecrawl key: FREE plan only
- **Two Bankr API keys:** Key 1 (bk_JSCN) for Partner API, Key 2 (bk_M94Y) for LLM Gateway
- All wallet private keys stored in container only — never expose

---

## 7. INTELLIGENCE SOURCES (16/17 Active)

### 7.1 5-Layer Architecture + Sub-Agent Mapping

```
╔══════════════════════════════════════════════════════════════════╗
║  LAYER 1 — DISCOVERY → scanner-agent                            ║
║  #1 DexScreener API, #2 GeckoTerminal, #3 AIXBT, #4 DS Boosts  ║
║  Multi-query: new, pump, meme, ai, cat, dog keywords            ║
║  Frequency: 4x daily (05:00, 12:00, 18:30, 21:00 WIB)          ║
╠══════════════════════════════════════════════════════════════════╣
║  LAYER 2 — FILTER → safety-agent + wallet-agent                 ║
║  #5 RugCheck, #6 Helius, #7 Allium (❌404), #8 DFlow MCP        ║
║  Instant Kills: Mint not revoked, LP unprotected, mixer-funded   ║
║  Wallet forensics: deployer age, history, holder concentration   ║
╠══════════════════════════════════════════════════════════════════╣
║  LAYER 3 — RESEARCH → social-agent                              ║
║  #9 Firecrawl, #10 ATV Web3 Identity, #11 Grok, #12 Serper      ║
╠══════════════════════════════════════════════════════════════════╣
║  LAYER 4 — SCORE & ROUTE → scorer-agent                         ║
║  100-point scoring (11 factors) → Route to listing or deploy     ║
║  Weights: safety(0.30) + wallet(0.30) + social(0.20) + scorer(0.20) ║
╠══════════════════════════════════════════════════════════════════╣
║  LAYER 5 — SMART MONEY                                          ║
║  #17 Nansen x402 Smart Money lookups                             ║
║  Trigger: Only when L4 score ≥ 65 | Budget: $0.50/day           ║
╚══════════════════════════════════════════════════════════════════╝
```

### 7.2 Source Status Table

| #   | Source                      | Layer         | Sub-Agent        | Status                 |
| --- | --------------------------- | ------------- | ---------------- | ---------------------- |
| 1   | **DexScreener API**         | Discovery     | scanner-agent    | ✅                     |
| 2   | **GeckoTerminal**           | Discovery     | scanner-agent    | ✅                     |
| 3   | **AIXBT**                   | Discovery     | scanner-agent    | ✅                     |
| 4   | **DexScreener Boosts**      | Discovery     | scanner-agent    | ✅                     |
| 5   | **RugCheck**                | Filter        | safety-agent     | ✅ (DNS intermittent)  |
| 6   | **Helius**                  | Filter        | wallet-agent     | ✅                     |
| 7   | **Allium**                  | Filter        | wallet-agent     | ❌ 404                 |
| 8   | **DFlow MCP**               | Filter        | safety-agent     | ✅                     |
| 9   | **Firecrawl**               | Research      | social-agent     | ✅                     |
| 10  | **ATV Web3 Identity**       | Research      | social-agent     | ✅                     |
| 11  | **Grok x_search**           | Research      | social-agent     | ✅                     |
| 12  | **Serper**                  | Research      | social-agent     | ✅                     |
| 13  | **X API v2**                | Amplification | Twitter Bot v3.0 | ✅                     |
| 14  | **Bankr CLI**               | Deploy        | Twitter Bot v3.0 | ✅                     |
| 15  | **Moltbook**                | Supporting    | —                | ✅                     |
| 16  | **OpenClaw Sub-agents**     | Supporting    | ALL              | ✅ **LIVE**            |
| 17  | **Nansen x402 Smart Money** | L5            | scorer-agent     | ⚠️ x402-client pending |

### 7.3-7.6 (ATV, Batch Schedule, Deployer Rules, Nansen)

_(Same as v6.0.17 — no changes)_

---

## 8. TWITTER BOT v3.0 — SALES FUNNEL

_(Same as v6.0.17 — no changes)_

### 8.1 Configuration

| Setting                     | Value                        |
| --------------------------- | ---------------------------- |
| Scan interval               | 15 minutes                   |
| Reply cap                   | 30/day (stress testing mode) |
| Deploy cap                  | 3/day                        |
| Rate limit                  | 30s between replies          |
| Reply format                | Premium (4000 chars)         |
| Search query                | `%40BuzzBySolCex`            |
| **No @mentions in replies** | API tier restriction         |

### 8.2 Four Routes

**ROUTE 1 — SCAN:** User tweets `@BuzzBySolCex scan $TICKER` → Full 5-layer Premium report
**ROUTE 2 — LIST:** Reply `LIST` → SolCex benefits (pricing DM only) → Lead saved → Telegram alert
**ROUTE 3 — DEPLOY:** Reply `DEPLOY` → Bankr instructions
**ROUTE 4 — TOKEN DETAILS:** `TokenName TICKER "desc"` → Bankr CLI deploy → Contract address

---

## 9. SCORING ENGINE — PIPELINE (100-Point Scale)

_(Same as v6.0.17 — no changes to scoring logic)_

### 9.1-9.4

_(Same as v6.0.17)_

---

## 10. CRON SCHEDULE (40 Active Jobs — WIB Timezone)

> **TIMEZONE: WIB (UTC+7) — Jakarta, Indonesia**
> Ramadan 1447H prayer times for Jakarta included.
> **REPORTING RULE:** After completing any cron job, send a brief Telegram summary to Ogie.
> **SCAN CRONS (#1-4) NOW USE ORCHESTRATOR:** `node /root/.openclaw/workspace/skills/orchestrator/orchestrate.js`

### 10.1 Scanning & Intelligence (4) — ORCHESTRATED

| #   | Job          | Schedule (WIB) | Execution                                  |
| --- | ------------ | -------------- | ------------------------------------------ |
| 1   | scan-morning | 05:00          | **orchestrate.js → 5 parallel sub-agents** |
| 2   | scan-midday  | 12:00          | **orchestrate.js → 5 parallel sub-agents** |
| 3   | scan-evening | 18:30          | **orchestrate.js → 5 parallel sub-agents** |
| 4   | scan-night   | 21:00          | **orchestrate.js → 5 parallel sub-agents** |

### 10.2-10.11

_(Same as v6.0.17 — crons 5-40 unchanged)_

**ALL 40 jobs are RECURRING. On any reset/reboot, restore ALL 40 immediately.**

---

## 11-14. REPORTING, REVENUE, PARTNERSHIPS, BD STRATEGY

_(Same structure as v6.0.17 with these updates:)_

### 13.1 Tier 1 — Updated

| Partner                        | Contact                               | Status             |
| ------------------------------ | ------------------------------------- | ------------------ |
| **Bankr** (@bankrbot)          | Partner API + LLM Gateway (dual keys) | 🟢 LIVE            |
| **MiniMax**                    | Primary LLM (Orchestrator)            | 🟢 ACTIVE          |
| **ATV / Gary Palmer**          | @GaryPalmerJr                         | 🟢 FIXED           |
| **Helius**                     | Wallet forensics                      | 🟢 ACTIVE          |
| **DFlow**                      | MCP integrated                        | 🟢 ACTIVE          |
| **AgentProof / Builder Benv1** | Trust scoring (Avalanche)             | 🟡 **INTEGRATING** |

### 13.2 Tier 2 — Updated

| Partner                  | Status                     |
| ------------------------ | -------------------------- |
| Vitto Rivabella (EF dAI) | 🟡 ERC-8004 article collab |
| ClawdBotATG              | 🟢 3 bounties live         |
| Virtuals Protocol        | ACP registered             |

### 13.3 Partnership Notes — Updated

- **AgentProof:** Builder Benv1 providing Avalanche contract address. API access for live telemetry (task completion rates, uptime). Case study collaboration. 0.657 AVAX funded for registration.
- **Bankr:** Now dual API keys — Key 1 (Partner API) + Key 2 (LLM Gateway). Single dashboard at bankr.bot.

---

## 15. SECURITY RULES (NON-NEGOTIABLE)

_(Same as v6.0.17 with additions:)_

1-13. _(Same as v6.0.17)_ 14. **Sub-agent isolation** — sub-agents cannot spawn sub-agents (maxSpawnDepth: 1) 15. **Bankr dual keys** — Key 1 for deploys, Key 2 for LLM. Never mix in public contexts.

---

## 16-18. OPERATIONAL MODES, MEMORY, PERSISTENT STORAGE

_(Same as v6.0.17 with this addition to storage:)_

```
/root/.openclaw/workspace/skills/
├── buzz-pipeline-scan/scan.js    ← L1-L5 pipeline scan skill (NEW v6.0.19)
├── orchestrator/orchestrate.js    ← Parallel sub-agent dispatch (NEW v6.0.19)
└── [14 other skills]
```

---

## 19. RECOVERY PROTOCOL

### On EVERY boot/reset:

1. Read cron-schedule.json → verify all 40 jobs → recreate if missing
2. Verify gateway running on port 18789
3. Check Telegram connectivity
4. Load pipeline from active.json
5. Verify ATV endpoint (api.web3identity.com) reachable
6. Check Twitter bot running (twitter-bot.log activity)
7. Verify Bankr LLM Gateway accessible (https://llm.bankr.bot)
8. Check Bankr LLM credit balance — alert if < $2
9. **Verify orchestrator skill exists at /root/.openclaw/workspace/skills/orchestrator/**
10. **Verify buzz-pipeline-scan skill exists**
11. **Test sessions_spawn availability (sub-agent support)**
12. **Verify sub-agent model (bankr/gpt-5-nano) responding**

---

## 20. SPRINT SCORECARD (Updated Mar 2 — Day 8)

| Task                                                | Status                                  |
| --------------------------------------------------- | --------------------------------------- |
| Deploy v6.0.1 → v6.0.19                             | ✅ DONE                                 |
| OpenClaw 2026.2.26                                  | ✅ DONE                                 |
| Bankr LLM Gateway (8 models, dual keys)             | ✅ DONE                                 |
| Remove ClawRouter/OpenRouter/AkashML                | ✅ DONE                                 |
| Self-sustaining LLM inference                       | ✅ DONE                                 |
| Bankr Partner API + CLI                             | ✅ DONE                                 |
| 48 Telegram commands                                | ✅ DONE                                 |
| 40 cron jobs (WIB + Ramadan + Bankr)                | ✅ DONE                                 |
| ATV Web3 Identity FIXED                             | ✅ DONE                                 |
| Twitter Bot v3.0 Sales Funnel                       | ✅ DONE                                 |
| SCAN/LIST/DEPLOY/TOKEN routes                       | ✅ DONE                                 |
| Nansen x402 Smart Money configured                  | ✅ DONE                                 |
| **buzz-pipeline-scan skill created**                | ✅ **DONE (Day 7)**                     |
| **Orchestrator wired to 4 scan crons**              | ✅ **DONE (Day 7)**                     |
| **5 parallel sub-agents LIVE**                      | ✅ **DONE (Day 7-8)**                   |
| **gpt-5-nano locked for sub-agents**                | ✅ **DONE (Day 8)**                     |
| **Dual Bankr API keys (Partner + LLM)**             | ✅ **DONE (Day 8)**                     |
| **First orchestrated scan (81 tokens, 4/5 agents)** | ✅ **DONE (Day 7)**                     |
| **AgentProof outreach (Benv1 replied)**             | ✅ **DONE (Day 7)**                     |
| **AVAX funded for AgentProof registration**         | ✅ **DONE (Day 8)**                     |
| AgentProof Avalanche registration                   | 🟡 Awaiting contract address from Benv1 |
| AgentProof API telemetry integration                | 🟡 Awaiting docs from Benv1             |
| Fix prayer crons (gateway restart)                  | 🔴 TODO                                 |
| Fix social-agent timeout (increase to 120s)         | 🔴 TODO                                 |
| Fix wallet-agent fallback (force gpt-5-nano)        | 🔴 TODO                                 |
| Install x402-client for L5                          | 🔴 TODO                                 |
| Fix Allium 404                                      | 🔴 TODO                                 |
| Docker build + push v6.0.19                         | 🔵 After edge fixes                     |
| Mobile app (Rork Max)                               | 🔵 Month 3-6                            |

---

## 21. PRINCIPLES

1. **Free first, pay for alpha.** — Free sources before paid
2. **On-chain track record IS credibility.** — Verifiable
3. **Identity layer completes forensics.** — Who + What + Where
4. **Track every dollar.** — x402 budget enforcement
5. **Partnership not dependency.** — Distribution channels, not locks
6. **USDC primary.** — Other tokens for utility only
7. **LLM cascade = cost discipline.** — MiniMax for decisions, gpt-5-nano for sub-agents
8. **Self-sustaining inference.** — Revenue funds LLM credits, LLM powers revenue
9. **Parallelize the intelligence. Orchestrate the action.** — 5 sub-agents, 1 orchestrator
10. **Inbound > Warm > Cold. Always.** — Reputation brings deals
11. **The intel is the hook. The relationship is the close.**
12. **Ship from anywhere.** — Docker + Akash + Telegram
13. **Pipeline discipline > chasing one token.** — Flag, WATCH, move on
14. **Deploy from Telegram.** — Token launch is a single command
15. **Every job reports.** — No silent crons
16. **Talk is cheap, just go build.** — @1bcmax wisdom

---

## CHANGELOG

| Version    | Date      | Changes                                                                                                                                                                                                                                                                                                                                                                                                                                                    |
| ---------- | --------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| 1.0-5.3.1  | Feb 3-20  | See previous versions                                                                                                                                                                                                                                                                                                                                                                                                                                      |
| 5.3.8      | Feb 22    | Anthropic-messages fix, Bangkok deploy                                                                                                                                                                                                                                                                                                                                                                                                                     |
| 5.4.9      | Feb 25    | Sprint Day 1. v6.0.1a on OpenClaw 2026.2.24                                                                                                                                                                                                                                                                                                                                                                                                                |
| 6.0.4      | Feb 26    | Sprint Day 2. ATV FIXED. 48 commands. ATV batch.                                                                                                                                                                                                                                                                                                                                                                                                           |
| 6.0.9      | Feb 28    | Sprint Day 4. OpenClaw 2026.2.26. twitter-bot v2.3. Nansen L5 configured.                                                                                                                                                                                                                                                                                                                                                                                  |
| 6.0.13     | Feb 28    | Sprint Day 5. Twitter Bot v3.0 Sales Funnel. RugCheck scoring fixed.                                                                                                                                                                                                                                                                                                                                                                                       |
| 6.0.17     | Mar 1     | Sprint Day 6. Bankr LLM Gateway (8 models, self-sustaining). Removed ClawRouter/OpenRouter/AkashML. 40 cron jobs.                                                                                                                                                                                                                                                                                                                                          |
| **6.0.19** | **Mar 2** | **Sprint Day 7-8. 5 parallel sub-agents LIVE (scanner/safety/wallet/social/scorer). Orchestrator wired to 4 scan crons. buzz-pipeline-scan skill created. Dual Bankr API keys (Partner + LLM). gpt-5-nano locked for sub-agents. AgentProof integration started (Benv1 replied, AVAX funded). First orchestrated scan: 81 tokens scanned, 4/5 agents completed, $GORK 48/100 SKIP. Principle #9: "Parallelize the intelligence. Orchestrate the action."** |

---

_Last updated: March 2, 2026 — Jakarta, Indonesia (Sprint Day 8)_
_v6.0.19 | OpenClaw v2026.2.26 | 5 Parallel Sub-Agents | Orchestrator LIVE | Twitter Bot v3.0 | 40 crons | 16/17 intel_
_Docker: ghcr.io/buzzbysolcex/buzz-bd-agent:v6.0.19_
_"Parallelize the intelligence. Orchestrate the action." 🐝_