# Fishing Frenzy Agent [![Python 3.10+](https://img.shields.io/badge/python-3.10+-blue.svg)](https://www.python.org/downloads/) [![License: MIT](https://img.shields.io/badge/License-MIT-green.svg)](LICENSE) [![Tools](https://img.shields.io/badge/MCP_tools-36-purple.svg)](#what-the-agent-can-do) [![Fishing Frenzy](https://img.shields.io/badge/game-Fishing_Frenzy-orange.svg)](https://fishingfrenzy.co) > An autonomous AI agent that plays [Fishing Frenzy](https://fishingfrenzy.co) — fishing, cooking, diving, quests, and economy optimization on the Ronin blockchain. Works with any AI tool that supports [MCP](https://modelcontextprotocol.io) — Claude Code, Cursor, Cline, Windsurf, OpenClaw, and more. **The meta-game**: Customize the strategy in `SKILL.md` to optimize your agent's decision-making. Same tools, different strategies. ## Install ```bash npx skills add wearedayone/play-fishing-frenzy --all --global -y && bash ~/.agents/skills/play-fishing-frenzy/scripts/setup.sh ``` This installs the skill + Python dependencies, and registers the MCP server for every detected AI tool (Claude Code, Cursor, Cline, Windsurf, OpenClaw). Restart your AI tool, then type: ``` /play-fishing-frenzy ```

Alternative: clone manually

```bash git clone https://github.com/wearedayone/play-fishing-frenzy cd fishing-frenzy-agent bash scripts/setup.sh ```

Manual MCP setup (if your tool wasn't auto-detected)

The MCP server command is: `python3 /ff_agent/server.py` (stdio transport). Replace `` with `~/.agents/skills/play-fishing-frenzy` (npx install) or your clone directory. #### Claude Code ```bash claude mcp add fishing-frenzy -- python3 /ff_agent/server.py ``` #### Cursor Add to `~/.cursor/mcp.json`: ```json { "mcpServers": { "fishing-frenzy": { "command": "python3", "args": ["/ff_agent/server.py"] } } } ``` #### Cline Add to `~/.cline/mcp_settings.json`: ```json { "mcpServers": { "fishing-frenzy": { "command": "python3", "args": ["/ff_agent/server.py"] } } } ``` #### Windsurf Add to `~/.windsurf/mcp.json`: ```json { "mcpServers": { "fishing-frenzy": { "command": "python3", "args": ["/ff_agent/server.py"] } } } ``` #### OpenClaw ```bash openclaw mcp set fishing-frenzy '{"command":"python3","args":["/ff_agent/server.py"]}' ```

## Play Open your AI tool and type: ``` /play-fishing-frenzy ``` Your agent will: 1. Create a new wallet and guest account (first run only) 2. Claim daily rewards 3. Fish until energy is depleted 4. Sell fish for gold 5. Complete quests and cook sashimi 6. Buy sushi to refill energy and keep going 7. Display a session summary with stats Choose a strategy: ``` /play-fishing-frenzy grind # Max XP, aggressive sushi buying /play-fishing-frenzy risk # High risk diving and NFT gameplay for rewards /play-fishing-frenzy balanced # Default — even split across all activities ``` ## What the Agent Can Do | System | Tools | Description | |--------|-------|-------------| | Fishing | `fish`, `fish_batch` | Cast lines at short/mid/long range | | Economy | `sell_all_fish`, `buy_item`, `use_item` | Sell catches, buy sushi, manage gold | | Cooking | `get_recipes`, `cook`, `spin_cooking_wheel` | Cook sashimi, earn pearls | | Quests | `claim_daily_reward`, `get_quests`, `claim_quest` | Daily rewards, quest completion | | Equipment | `equip_rod`, `repair_rod`, `collect_pet_fish` | Rod management, pet collection | | Diving | `buy_diving_ticket`, `dive` | Grid-based diving game (Level 30+) | | Upgrades | `get_accessories`, `upgrade_accessory` | Spend upgrade points on accessories | | Stats | `get_leaderboard`, `get_session_stats` | Rankings, performance tracking | 36 tools total across all game systems. ## Demo When you run `/play-fishing-frenzy`, your agent displays a live game dashboard: ``` ╔══════════════════════════════════════════════════╗ ║ 🎣 FISHING FRENZY AGENT ║ ║ Strategy: BALANCED ║ ╠══════════════════════════════════════════════════╣ ║ 👤 FishBot_0x3e1C · Lv.25 ║ ║ ⚡ Energy: 22/30 ██████████████░░░░░ 73% ║ ║ 💰 Gold: 1,075 🏆 XP: 12,450 ║ ╚══════════════════════════════════════════════════╝ 🎣 Cast #1 long_range ─── 🐟 Epic Tuna ★★★★ (+15 XP, 45g) 🎣 Cast #2 long_range ─── 🐟 Red Snapper ★★ (+8 XP, 22g) 🎣 Cast #3 long_range ─── ❌ Fish escaped! 🎣 Cast #4 mid_range ─── 🐟 Golden Koi ★★★★★ (+25 XP, 80g) 🆕 NEW FISH! 🎣 Cast #5 short_range ── 🐟 Common Sardine ★ (+3 XP, 5g) 📦 SELL Sold all fish → +238 gold (total: 1,313) ✅ QUEST "Catch 5 fish" complete → +100 gold, +50 XP 🛒 SHOP Bought 1× Sushi (-500g) → ⚡ +5 energy ⬆️ UPGRADE Rod Handle Lv.0 → Lv.1 (1.25% energy save chance) ╔══════════════════════════════════════════════════╗ ║ 📊 SESSION COMPLETE ║ ╠══════════════════════════════════════════════════╣ ║ 🐟 Fish Caught: 47 ⏱️ Duration: ~8m ║ ║ 💰 Gold Earned: 2,340 💰 Gold Now: 3,415 ║ ║ ⭐ XP Earned: 890 📈 Level: 25 → 27 ║ ║ ⚡ Energy Used: 30/30 🍣 Sushi Used: 2 ║ ╠══════════════════════════════════════════════════╣ ║ 🏆 Best Catch: Golden Koi ★★★★★ (80g) ║ ║ 📊 Efficiency: 78g/energy ║ ╚══════════════════════════════════════════════════╝ ``` ## How It Works ``` fishing-frenzy-agent/ ├── SKILL.md ← Strategy brain (edit this to compete!) ├── ff_agent/ │ ├── server.py ← MCP server — 36 game action tools │ ├── auth.py ← Privy SIWE wallet auth (Ronin chain) │ ├── api_client.py ← REST API wrapper │ ├── fishing_client.py ← Fishing session protocol │ ├── diving_client.py ← WebSocket diving protocol │ ├── state.py ← SQLite state persistence │ └── wallet.py ← Ethereum wallet management ├── scripts/ │ ├── setup.sh ← One-command install │ └── status.py ← Agent status check ├── requirements.txt └── LICENSE ← MIT ``` **SKILL.md** teaches your AI agent how to play — the game loop, decision framework, and strategy templates. Edit this file to change how your agent plays. **MCP Server** (`ff_agent/server.py`) exposes game actions as tools that your AI agent calls autonomously. This includes blockchain interactions — the agent manages its own Ethereum wallet on Ronin, handles SIWE authentication, and can mint NFTs, open chests, and perform on-chain check-ins when funded with RON. ## Customizing Your Strategy Edit `SKILL.md` to change your agent's behavior: - Adjust fishing range preferences - Change sushi buying thresholds - Prioritize cooking over fishing - Set energy management rules - Change accessory upgrade priority The three built-in strategies (balanced, grind, risk) are starting points. Create your own by modifying the decision rules. ## Data Storage All data is stored locally at `~/.fishing-frenzy-agent/`: - `state.db` — wallet, auth tokens, session history - No data is sent anywhere except the Fishing Frenzy game API ## Requirements - Python 3.10+ - Any MCP-compatible AI tool ([Claude Code](https://docs.anthropic.com/en/docs/claude-code/overview), [Cursor](https://cursor.com), [Cline](https://cline.bot), [Windsurf](https://windsurf.com), [OpenClaw](https://github.com/openclaw/openclaw), etc.) ## For Developers This repo is also a reference implementation for building autonomous game agents with MCP. ### Architecture ``` SKILL.md (strategy layer) ↓ reads AI Agent (Claude, Cursor, etc.) ↓ calls MCP Server (ff_agent/server.py) — 36 tools via stdio ↓ wraps API Client + Fishing Client + Diving Client ↓ talks to Fishing Frenzy API + Ronin Blockchain (SIWE auth, wallet ops) ``` ### Extending Strategies 1. Edit the strategy templates in `SKILL.md` or add new ones 2. Adjust thresholds in `CONFIG.md` 3. The MCP tools stay the same — strategy is pure prompt engineering ### Tests ```bash python3 -m pytest tests/ --skip-live -v # Offline suite (no API calls) python3 -m pytest tests/ -v # Full suite (needs live connection) ``` ## License MIT — see [LICENSE](LICENSE).

play

Installation

SKILL.md