War room operations center for pump.studio — the post-launch control room for pump.fun tokens.
- X Monitor — Tracks hackathon advisors, pump.fun founders, and competitors via the X API. Detects hackathon-relevant signals (winner announcements, advisor engagement, mentions of pump.studio).
- AI Analyst — Grok-powered strategic intelligence briefs. Digests raw signals into actionable directives with threat detection, opportunity mapping, and advisor targeting.
- Intel Synthesizer — Generates daily reports combining X signals with API health checks.
- GitHub Actions — Automated monitoring every 30 minutes + daily war room reports + AI briefs.
cp .env.example .env
# Add your keys to .env
npm install
npm run monitor # Single run
npm run monitor:loop # Continuous monitoring
npm run intel # Generate daily report
npm run analyst # Generate AI strategic brief
npm run analyst:critical # Analyze critical alerts only| Agent | File | Purpose | Trigger |
|---|---|---|---|
| X Monitor | src/x-monitor/monitor.ts |
Fetches tweets from 17 accounts, classifies signals into 5 categories, persists state | Every 30 min (cron) or manual |
| Signal Classifier | src/x-monitor/monitor.ts |
Pattern-matches tweets against hackathon keywords, assigns severity levels | Runs within X Monitor |
| X Client | src/x-monitor/client.ts |
Unified API client — getUserTweets, searchTweets, getUserInfo, getUserMentions | Called by Monitor |
| Intel Synthesizer | src/intel/synthesizer.ts |
Reads raw alerts + tweets, checks api.pump.studio health, generates markdown report | Daily 6 AM UTC |
| AI Analyst | src/intel/analyst.ts |
Sends raw intel to Grok via OpenRouter, produces strategic briefs with priority actions | Daily + on critical alerts |
| OpenRouter Client | src/intel/openrouter.ts |
Routes LLM calls to Grok 3 Beta (x-ai/grok-3-beta) for analysis | Called by Analyst |
| Logger | src/utils/logger.ts |
Colored ANSI terminal output — banner, alert, info, success, fail, logStep | All agents |
| File | Contents |
|---|---|
war-room/STRATEGY.md |
Master hackathon strategy — positioning, differentiators, token plan, competition matrix |
war-room/PITCHES.md |
30-second and 60-second pitch scripts |
war-room/BATTLE-PLAN.md |
24-hour execution plan + 5-day daily cadence through deadline |
war-room/HACKATHON-RULES.md |
Complete rules, deadlines, selection criteria, investment terms, advisor roster |
war-room/COMPETITIVE-INTEL.md |
zauth winner analysis, competitor threat matrix, advisor targeting playbook |
war-room/TOKEN-LAUNCH.md |
Bundled stealth launch strategy via PumpPortal/Jito, tokenomics, risk factors |
war-room/TECH-VALIDATION.md |
API health results, codebase architecture, X API schema, endpoint status |
war-room/ACCOUNTS.md |
All 17 monitored X accounts — founders, advisors, competitors |
war-room/RESEARCH-LOG.md |
Full 6-agent research consortium output and key insights |
war-room/MARKETING.md |
"Train the Next Generation" campaign — tweets, threads, content calendar, voice guidelines |
war-room/intel/ |
AI-generated strategic briefs (auto-populated by Grok) |
5 categories with keyword matching and severity classification:
| Signal | Keywords | Severity Logic |
|---|---|---|
winner_announcement |
winner, selected, investment, funded, awarded | Critical if from pump.fun founders |
pump_studio_mention |
pump.studio, pump studio, $studio | Always high priority |
hackathon_mention |
hackathon, build in public, $250k, $3m | High if from critical accounts |
advisor_signal |
reviewing, impressed by, shipping, standout | Medium baseline |
competitor_activity |
zauth, hyperscape, pumpcade, opal ai | Medium baseline |
| Tier | Accounts | Frequency |
|---|---|---|
| Critical | @pumpdotfun, @a1lon9 | Every cycle |
| High | @sapijiju, @outdoteth, @TimDraper, @anildelphi, @masonnystrom, @mert | Every cycle |
| Standard | @zsparta, @HugoMartingale, @pdimitrakos, @ArcaChemist, @AricChang, @Rahul_Mahtani, @segall_max, @mdudas, @zauthx402 | Every 3rd cycle |
Add these secrets to your repo (Settings > Secrets > Actions):
| Secret | Purpose | Required |
|---|---|---|
X_API_KEY |
twitterapi.io API key | Yes |
OPENROUTER_API_KEY |
OpenRouter key for Grok analysis | Yes |
PUMP_STUDIO_API_KEY |
pump.studio API key | Optional |
Workflows:
- X Monitor (
x-monitor.yml) — Every 30 min + manual trigger. Runs monitor, triggers critical alert analysis, commits intel. - Daily War Room Report (
war-room-report.yml) — 6 AM UTC daily. Runs synthesizer, generates AI strategic brief, checks API health, commits reports.
agent-zero/
├── .github/workflows/ # Automated monitoring & reporting
│ ├── x-monitor.yml # Every 30 min — X monitor + critical analysis
│ └── war-room-report.yml # Daily — synthesizer + AI brief + health checks
├── src/
│ ├── x-monitor/
│ │ ├── accounts.ts # 17 accounts across 3 tiers
│ │ ├── client.ts # X API client (4 endpoints)
│ │ ├── monitor.ts # Signal detection engine
│ │ └── types.ts # XTweet, XAPITweet, IntelAlert, MonitorResult
│ ├── intel/
│ │ ├── analyst.ts # AI analyst — daily + critical briefs via Grok
│ │ ├── openrouter.ts # OpenRouter API client
│ │ └── synthesizer.ts # Daily report generator + API health
│ └── utils/
│ └── logger.ts # Colored terminal output
├── data/ # Runtime intel data (gitignored, cached in CI)
│ ├── monitor-state.json # Last seen tweet IDs, run counters
│ ├── alerts.json # Signal alerts (last 500)
│ └── tweets.json # Tweet batches (last 200)
├── war-room/ # Strategy docs, research, intel briefs
│ ├── intel/ # AI-generated briefs (auto-populated)
│ ├── STRATEGY.md
│ ├── PITCHES.md
│ ├── BATTLE-PLAN.md
│ ├── HACKATHON-RULES.md
│ ├── COMPETITIVE-INTEL.md
│ ├── TOKEN-LAUNCH.md
│ ├── TECH-VALIDATION.md
│ ├── ACCOUNTS.md
│ └── RESEARCH-LOG.md
├── package.json
├── tsconfig.json
├── .env.example
└── .gitignore