bastion-mcp-server
Risk Intelligence MCP Server for Crypto Agents | 52 tools, 560+ signals, 72B AI model
Install / Use
claude mcp add bastionfintech-spec -- npx -y github:bastionfintech-spec/bastion-mcp-serverIf the server publishes to npm under a different name, use that package instead — check the repo README.
MCP Server
Model Context Protocol server
Quality Score
Category
AI & Machine LearningSupported Platforms
Tags
Our assessment of bastion-mcp-server
bastion-mcp-server scores 75/100 on our quality scale, 679th of 822 AI & Machine Learning skills we index.
Its MCP Server is 33 KB long, well organised into 33 sections with 20 code examples: a thorough specification that gives an agent plenty to work with.
It has 3 GitHub stars, so there is little community track record yet; judge it on its content.
Maintenance, license and trust
- The repository was last updated about 7 months ago. That is recent enough to be usable, but agent tooling moves fast, so check the instructions against your agent's current version.
- Our last check on 2026-09-20 found the source still online.
- It is released under the MIT license, a permissive license that allows use, modification and commercial use with attribution.
- Its trust signals score 81/100, with 3 cautions from licensing, adoption, age or documentation. These come from repository metadata, not a code audit — read the skill file before letting an agent act on it.
bastion-mcp-server compared with similar skills
All 4 of these similar skills score higher than bastion-mcp-server; compare them before choosing.
| Skill | Score | Stars | Updated | Format |
|---|---|---|---|---|
| bastion-mcp-server (this skill)by bastionfintech-spec | 75 | 3 | 7mo ago | MCP Server |
| claude-memby thedotmack | 100 | 94.9k | today | CLAUDE.md |
| Agent-Reachby Panniantong | 100 | 86.1k | 13d ago | CLAUDE.md |
| Understand-Anythingby Egonex-AI | 100 | 84.6k | 1d ago | CLAUDE.md |
| headroomby headroomlabs-ai | 100 | 74.1k | today | CLAUDE.md |
Frequently asked questions
- How do I install bastion-mcp-server?
- Run
claude mcp add bastionfintech-spec -- npx -y github:bastionfintech-spec/bastion-mcp-server. The install tabs above show the steps for each supported agent. - Which AI agents does bastion-mcp-server work with?
- It is written for Claude Code and Claude Desktop, as a MCP Server file. Other agents that read the same format can often use it too.
- Is bastion-mcp-server safe to use?
- It is MIT-licensed and scores 81/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
- Is bastion-mcp-server still maintained?
- The repository was last updated about 7 months ago. That is recent enough to be usable, but agent tooling moves fast, so check the instructions against your agent's current version.
Skill content
View source on GitHubBASTION
Risk Intelligence MCP Server for Crypto Agents
<br />75.4% Win Rate | 147 Tools | 560+ Signals | 72B Parameter AI | 5 Exchanges
<br /> <br />Quick Start · Features · Architecture · Examples · Docs
<br />Platform • Agents Hub • Developers • Get API Key
</div> <br /><br />[!IMPORTANT] Zero install required. BASTION is a cloud-hosted MCP server. Connect with one command and your Claude agent instantly gets 147 real-time crypto tools. No dependencies, no Docker, no API wrappers.
What is BASTION?
BASTION is a Model Context Protocol (MCP) server that turns any Claude agent into a professional crypto trading desk. Through a single MCP connection (Streamable HTTP or SSE), your agent gains 147 specialized tools for market data, derivatives analytics, on-chain intelligence, AI risk evaluation, portfolio management, and autonomous trade execution.
The core engine is a fine-tuned 72B parameter model running on a 4x RTX 5090 GPU cluster, analyzing 560+ real-time signals per evaluation. It doesn't just return data — it returns structured risk assessments with reasoning chains, confidence scores, and explicit action recommendations.
You: "Check my BTC long, I'm worried about the funding spike"
Claude: [Calls bastion_evaluate_risk, bastion_get_funding_rates, bastion_get_liquidations]
Your BTC LONG is at moderate risk (52/100). Here's why:
Funding at 0.035% is elevated but not extreme. However, there's a
$2.1B liquidation cluster at $94,200 — just 3.8% below current price.
OI rose 4.2% without corresponding volume, suggesting overleveraged
longs. CVD is diverging bearish on the 1h.
Recommendation: TP_PARTIAL — Take 50% profit here. Move stop to
breakeven on the remainder.
<br />[!TIP] All market data, derivatives, on-chain, macro, and AI tools work without an API key. You only need a key for portfolio access and trade execution.
Quick Start
Claude Code — One Command
# Streamable HTTP (recommended)
claude mcp add bastion-mcp --transport http https://bastionfi.tech/mcp/stream
# Or legacy SSE
claude mcp add bastion-mcp --transport sse https://bastionfi.tech/mcp/sse
That's it. Restart Claude Code. All 147 tools are live.
Claude Desktop — One Config
Add to claude_desktop_config.json (<kbd>macOS</kbd> ~/Library/Application Support/Claude/ · <kbd>Windows</kbd> %APPDATA%\Claude\):
{
"mcpServers": {
"bastion": {
"type": "streamable-http",
"url": "https://bastionfi.tech/mcp/stream"
}
}
}
<details>
<summary>Or use legacy SSE transport</summary>
{
"mcpServers": {
"bastion": {
"type": "sse",
"url": "https://bastionfi.tech/mcp/sse"
}
}
}
</details>
Python MCP Client
from mcp import ClientSession
from mcp.client.streamable_http import streamablehttp_client
# Streamable HTTP (recommended)
async with streamablehttp_client("https://bastionfi.tech/mcp/stream") as (read, write, _):
async with ClientSession(read, write) as session:
await session.initialize()
result = await session.call_tool("bastion_evaluate_risk", {
"symbol": "BTC",
"direction": "LONG",
"entry_price": 96200.0,
"current_price": 98400.0,
"leverage": 5,
"stop_loss": 94000.0
})
<details>
<summary>Legacy SSE client</summary>
from mcp import ClientSession
from mcp.client.sse import sse_client
async with sse_client("https://bastionfi.tech/mcp/sse") as (read, write):
async with ClientSession(read, write) as session:
await session.initialize()
result = await session.call_tool("bastion_evaluate_risk", {...})
</details>
Authenticated Access
For portfolio, trading, and engine tools — pass your key in the URL:
https://bastionfi.tech/mcp/stream?api_key=bst_…[redacted]
Get a key at bastionfi.tech/account | Full walkthrough in quickstart.md | Detailed guides in guides/
Features
Key Capabilities
| Capability | Description | |:---|:---| | Risk Intelligence | 560+ signal evaluation with reasoning chains and action recommendations | | Derivatives Analytics | OI, CVD, funding, liquidation heatmaps, options flow, gamma exposure | | On-Chain Intelligence | Whale tracking, exchange flow, accumulation detection | | Macro & Sentiment | DXY, ETF flows, Fear/Greed, Polymarket, FRED economic data | | Market Data | Real-time price, OHLCV klines, volatility metrics | | Portfolio Management | Multi-exchange positions, balances, alerts, session stats | | Trade Execution | Emergency exits, partial closes, TP/SL, breakeven moves | | Autonomous Engine | Self-monitoring risk engine with armed/disarmed modes | | Research Reports | AI-generated deep analysis reports with position sizing |
147 Tools Across 35 Categories
<details open> <summary><strong>Core AI — 4 tools</strong></summary>| Tool | Description |
|:-----|:------------|
| bastion_evaluate_risk | Full risk evaluation with 560+ signals, reasoning chain, and action recommendation |
| bastion_chat | Natural language conversation with the fine-tuned 72B risk model |
| bastion_evaluate_all_positions | Batch risk evaluation of every open position |
| bastion_scan_signals | Proactive signal scanner — finds trade opportunities ranked by conviction |
| Tool | Description |
|:-----|:------------|
| bastion_get_price | Real-time price, 24h change, volume |
| bastion_get_market_data | Market cap, circulating supply, dominance, ATH tracking |
| bastion_get_klines | OHLCV candlestick data — any timeframe from 1m to 1M |
| bastion_get_volatility | Realized volatility, ATR, Bollinger bandwidth, IV percentile |
| Tool | Description |
|:-----|:------------|
| bastion_get_open_interest | Aggregate open interest across all exchanges |
| bastion_get_oi_changes | OI delta over time windows (confirms or denies price moves) |
| bastion_get_cvd | Cumulative Volume Delta — buyer vs seller aggression |
| bastion_get_orderflow | Order flow imbalance and absorption analysis |
| bastion_get_funding_rates | Perpetual funding rates across 5 exchanges |
| bastion_get_funding_arb | Cross-exchange funding rate arbitrage opportunities |
| bastion_get_liquidations | Real-time and aggregate liquidation data |
| bastion_get_heatmap | Liquidation heatmap — see where cascading liquidations cluster |
| bastion_get_taker_ratio | Taker buy/sell ratio (real-time directional aggression) |
| bastion_get_top_traders | Top trader long/short positioning ratios |
| bastion_get_market_maker_magnet | Market maker gamma exposure and magnet price levels |
| bastion_get_options | Options flow, max pain, put/call ratio, IV skew |
| Tool | Description |
|:-----|:------------|
| bastion_get_whale_activity | Large wallet movements, accumulation/distribution patterns |
| bastion_get_exchange_flow | Exchange inflow/outflow trends (deposit vs withdrawal) |
| bastion_get_onchain | Active addresses, NVT ratio, SOPR, network metrics |
| bastion_get_news | AI-curated crypto news feed with sentiment scoring |
| Tool | Description |
|:-----|:------------|
| bastion_get_fear_greed | Crypto Fear & Greed Index with historical context |
| bastion_get_macro_signals | DXY, US10Y, SPX correlation, macro regime detection |
| bastion_get_etf_flows | Bitcoin & Ethereum spot ETF daily flow data |
| bastion_get_stablecoin_markets | Stablecoin supply changes and capital flow analysis |
| bastion_get_economic_data | CPI, FOMC, NFP, unemployment — via FRED API |
| bastion_get_polymarket | Prediction market odds for crypto-relevant events |
| Tool | Auth | Description |
|:-----|:----:|:------------|
| bastion_get_positions | read | All open positions across connected exchanges |
| bastion_get_balance | read | Account balance and margin information |
| bastion_get_exchanges | read | Connected exchange accounts and status |
| bastion_engine_status | read | Autonomous engine status and configuration |
| bastion_engine_history | read | Engine action history and audit log |
| bastion_get_alerts | read | Active alerts and notifications |
| bastion_get_session_stats | read | Trading session performance statistics |
| Tool | Auth | Reversible | Description |
|:-----|:----:|:----------:|:------------|
| bastion_emergency_exit | trade | No | Close position at market immediately |
| bastion_partial_close | trade | No | Close a percentage of a position |
| bastion_set_take_profit | trade | Yes | Set or modify take-profit order |
| bastion_set_stop_loss | trade | Yes | Set or modify stop-loss order |
| bastion_move_to_breakeven | trade | Yes | Move stop-loss to entry price |
| bastion_flatten_winners | trade | No | Close all positions currently in profit |
| Tool | Auth | Description |
|:-----|:----:|:------------|
| bastion_engine_start | engine | Start the autonomous risk engine |
| bastion_engine_arm | engine | Arm for live trade execution |
| bastion_engine_disarm | engine | Disarm to monitor-only mode |
STOPPED → engine_start → MONITORING → engine_arm → ARMED (live)
│
engine_disarm
│
MONITORING
</details>
<details>
<summary><strong>WTruncated for display — read the full file on GitHub.
Related Skills
claude-mem
94.9kPersistent Context Across Sessions for Every Agent – Captures everything your agent does during sessions, compresses it with AI, and injects relevant context back into future sessions. Works with Claude Code, OpenClaw, Codex, Gemini, Hermes, Copilot, OpenCode + More
Agent-Reach
86.1kGive your AI agent eyes to see the entire internet. Read & search Twitter, Reddit, YouTube, GitHub, Bilibili, XiaoHongShu — one CLI, zero API fees.
Understand-Anything
84.6kGraphs that teach > graphs that impress. Turn any code into an interactive knowledge graph you can explore, search, and ask questions about. Works with Claude Code, Codex, Cursor, Copilot, Gemini CLI, and more.
headroom
74.1kCompress tool outputs, logs, files, and RAG chunks before they reach the LLM. 20% fewer tokens for coding agents, 60-95% fewer tokens for JSON, same answers. Library, proxy, MCP server.
Trust signals
From repository metadata: license, adoption, age and documentation. Not a code audit — see the Safety scan above for what the skill file itself contains.
