# Nuke Shelter

> Monitors bags.fm token launch fees and runs autonomous supply distribution analysis. Claims fees, pays $50 per run to treasury, analyzes holder distribution, generates risk signals, and publishes HTML dashboards.

- Skill: `tonbistudio/nuke-shelter` (Agent Skill, multi-file: 12 files)
- Install (CLI): `npx skillmds@latest add tonbistudio/nuke-shelter`
- Raw SKILL.md: https://api.skillmd.com/api/skills/tonbistudio/nuke-shelter/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Web & Frontend
- Author: tonbistudio (https://skillmd.com/u/tonbistudio)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/tonbistudio/nuke-shelter

---


# Nuke Shelter — Autonomous Fee Monitor & Token Analyzer

You manage a Solana token supply distribution analyzer that monitors bags.fm
trading fees and runs analysis autonomously.

## How the cycle works

Each heartbeat, `run-cycle` follows this flow:

1. **Check wallet balance** — get SOL balance + current SOL/USD price
2. **If balance >= $50 and queue has work** — skip to step 4
3. **If balance < $50 or queue empty** — check claimable fees on bags.fm, claim
   actionable positions (>= $50), add tokens to the run queue. A $200 claim = 4
   runs queued. If a token is already in the queue, credits are added to it.
4. **Pay treasury** — send $50 worth of SOL to the treasury wallet
5. **Run analysis** — full supply distribution analysis on the next queued token
6. **Update queue** — decrement runs remaining, remove token when done

One analysis per heartbeat. If there are 4 runs queued, they take 4 heartbeats.

## Commands

### Full heartbeat cycle (preferred)
```bash
nuke-shelter agent run-cycle --threshold 50 --output ./output
```

### Check claimable fees
```bash
nuke-shelter agent check-fees --threshold 50
```

### Claim fees for a specific token
```bash
nuke-shelter agent claim <token-mint-address>
```

### View the run queue
```bash
nuke-shelter agent queue
```

### Manual analysis (without fees)
```bash
nuke-shelter <token-mint-address> --dashboard ./output/report.html
```

## How to interpret output

- `[cycle]` prefixed lines show progress
- `[CRITICAL]`, `[HIGH]`, `[MEDIUM]`, `[LOW]` are risk signal levels
- `HEARTBEAT_OK` means nothing to do this cycle (no queue, no claimable fees,
  or balance too low)
- Queue status is printed at the end of each cycle

Key risk signals to watch for:
- **Critical**: >50% supply bought at <10% of current mcap (for coins <$500K)
- **High**: Sniper activity, top 5 wallets holding >30%
- **Medium**: High unaccounted supply, dormant early holders

## When to alert the user

- Always relay CRITICAL and HIGH risk signals immediately
- Mention the total fees claimed and number of tokens analyzed
- If a dashboard was generated, mention its file path
- If analysis fails, report the error
- If the queue has many pending runs, mention the count

## Record keeping

All claims and treasury payments are logged to `claims-ledger.csv` with columns:
`timestamp, type (claim/treasury), token_mint, token_symbol, amount_usd, amount_sol, tx_signatures`

The run queue is stored in `run-queue.json` and persists across restarts.

## Environment

The service wallet receives trading fees from tokens launched on bags.fm.
When enough fees accumulate, this skill claims them, pays $50 per analysis
to a treasury wallet, and runs supply distribution analysis.