# io.github.TeleKashOracle/mcp-server (npm · telekash-mcp-server)

Prediction market probability oracle. 500+ markets from Kalshi, Polymarket & Metaculus.

- Trust score: 68/100 (medium)
- Change this week: +25
- Registry status: active
- Liveness: live
- Owner verified: no
- Last scored: 2026-08-03

## Components

- npm · `telekash-mcp-server`: 68/100 (this document), [markdown](https://verifymcp.io/servers/telekashoracle-mcp-server/telekash-mcp-server.md), [page](https://verifymcp.io/servers/telekashoracle-mcp-server/telekash-mcp-server)

## Channel facts

- Registry: `npm`
- Package: `telekash-mcp-server`
- Version: `0.6.1`
- Transport: `stdio`

## Trust breakdown

How this component scores in each security and reliability category. Every signal is checked automatically from public evidence about the published package, including repeated runs of it in an isolated sandbox, and we only credit what we can confirm. Scores are 0–100 per category. Scoring method: https://verifymcp.io/docs/scoring (what has changed: https://verifymcp.io/docs/scoring/changelog)

Scored 2026-08-03.

- **Supply Chain Security**: 87/100
  - No malware found by supply-chain analysis.
  - Only part of the dependency tree could be resolved (104 of 108), so this covers what we could see, not the whole tree.
  - No install/post-install scripts declared.
  - Only part of the dependency tree could be resolved (104 of 108), so this covers what we could see, not the whole tree.
- **Provenance & Transparency**: 45/100
  - Source repository is publicly reachable at the declared URL.
  - Provenance check failed: no build-provenance attestation is published.
  - Clear OSI-approved license (MIT).
  - Actively maintained (last published 135 days ago).
  - Disclosure check failed: no security disclosure policy was found in the source repository.
- **Schema Quality & AI Usability**: 70/100
  - AI-judged instruction clarity (excellent).
  - Context-footprint check failed: tool/resource definitions use about 2629 tokens (~175/item across 15 items; 15 tools + 0 resources), over budget; trim descriptions and params.
  - Usage-examples check failed: none of the tools include examples.
- **Stability & Change Management**: 27/100
  - Stability observed for 8 of 30 days with no destabilising changes; credit accrues until the full window elapses.
- **Tool Coverage**: 100/100
  - 100% of tools have a non-trivial description (not blank, and not just the tool's name).
  - 100% of tool parameters carry a description.
- **Capabilities**: 100/100
  - Implements a supported MCP spec version (2025-11-25); the latest is 2026-07-28.

## Install

### Claude

```bash
claude mcp add telekashoracle-mcp-server -- npx -y telekash-mcp-server
```

### Codex

```bash
codex mcp add telekashoracle-mcp-server -- npx -y telekash-mcp-server
```

### opencode

```json
{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "telekashoracle-mcp-server": {
      "type": "local",
      "command": [
        "npx",
        "-y",
        "telekash-mcp-server"
      ],
      "enabled": true
    }
  }
}
```

### OpenClaw

```bash
openclaw mcp add telekashoracle-mcp-server --command npx --arg -y --arg telekash-mcp-server
```

### Hermes

```yaml
mcp_servers:
  telekashoracle-mcp-server:
    command: "npx"
    args: ["-y", "telekash-mcp-server"]
```

### Other

```json
{
  "mcpServers": {
    "telekashoracle-mcp-server": {
      "command": "npx",
      "args": [
        "-y",
        "telekash-mcp-server"
      ]
    }
  }
}
```

## Changelog

Every change recorded for this component, newest first. Days that predate change tracking, or that we cannot explain, say so: "we were watching and nothing happened" and "we were not watching" are different claims.

### 2026-08-02 (score 68, +49)

- [security regression] Provenance: unverified → fail
- [security improvement] Install scripts: unverified → pass
- [security improvement] Known CVEs: unverified → partial
- [security improvement] Malware scan: unverified → pass
- [functional improvement] License: unverified → pass
- [functional improvement] Dependency health: unverified → partial
- [functional improvement] Maintenance: unverified → pass
- [functional improvement] MCP protocol: unverified → pass
- [functional improvement] Stability: unverified → 0.23
- [functional improvement] Schema quality: unverified → excellent
- [functional] Licence: MIT

### 2026-08-01 (score 19, +14)

- [functional improvement] Tool coverage: unverified → 100

### 2026-07-31 (score 5, −38)

- [functional] We updated how we score, so this day's move reflects our rubric, not a change to the server

### 2026-07-27 (score 43)

First indexed and scored.

## MCP tools (15)

### `get_probability` (~204 tokens)

Get real-time probability, odds, and likelihood for any prediction market outcome.

Returns YES/NO probabilities (0-100%), trading volume, liquidity depth, and market metadata from Kalshi (CFTC-regulated) and Polymarket.
Use this when asked about chances, odds, likelihood, forecasts, or predictions for any event — elections, crypto prices, sports, economics, weather, entertainment.

Example queries:
\- "What are the odds Trump wins 2028?" → election forecasting
\- "What's the probability BTC hits $200K?" → crypto price prediction
\- "Will the Fed cut rates?" → economic forecasting, interest rates
\- "What's the chance of rain in NYC?" → weather betting
\- "Who will win the Super Bowl?" → sports odds

Input parameters:

- `market_id` (string): The market UUID or external_id (ticker) to query
- `query` (string): Natural language query to search for a market (alternative to market_id)

### `list_markets` (~227 tokens)

Browse and discover prediction markets across 7 categories with filtering and sorting.

Lists active betting markets from Kalshi, Polymarket, and Metaculus. Filter by category, sort by trading volume, probability, or closing date. 500+ markets available.
Categories: sports, crypto, politics, economics, pop_culture, weather, other.
Use when exploring what predictions are available, finding trending markets, or discovering betting opportunities.

Example queries:
\- "Show me crypto prediction markets" → Bitcoin, Ethereum, altcoin forecasts
\- "What sports markets are trending?" → NFL, NBA, soccer odds
\- "List political predictions" → elections, legislation, geopolitics
\- "What economic forecasts are available?" → GDP, inflation, interest rates

Input parameters:

- `category` (string): Filter by category (default: all)
- `limit` (number): Maximum markets to return (default: 10, max: 50)
- `sort_by` (string): Sort order (default: volume)
- `source` (string): Filter by data source (default: all)

### `get_history` (~120 tokens)

Get historical probability changes and trend data for a prediction market over time.

Returns probability snapshots showing how odds, sentiment, and market consensus have shifted over 1h, 24h, 7d, or 30d.
Use for trend analysis, momentum detection, volatility assessment, and understanding how predictions evolve.
Essential for backtesting strategies, identifying probability swings, and spotting market-moving events.

Input parameters:

- `market_id` (string, required): The market UUID or external_id (ticker)
- `timeframe` (string): Time range for history (default: 24h)

### `search_markets` (~192 tokens)

Search 500+ prediction markets by keyword, topic, or natural language query.

Full-text search across Kalshi, Polymarket, and Metaculus. Finds markets matching any topic — politics, crypto, sports, economics, entertainment, science, technology, weather.
Returns matching active markets sorted by relevance and trading volume.
Use when looking for specific predictions, events, or outcomes to bet on.

Example queries:
\- "Trump election 2028" → presidential race odds
\- "Bitcoin price prediction" → BTC price target markets
\- "Super Bowl winner" → NFL championship odds
\- "AI regulation" → technology policy predictions
\- "Fed interest rate" → monetary policy forecasts

Input parameters:

- `limit` (number): Maximum results to return (default: 10, max: 50)
- `query` (string, required): Search query (e.g., 'Trump', 'Bitcoin', 'Super Bowl')

### `get_sentiment` (~104 tokens)

Get AI-powered sentiment analysis, recommendation, and confidence score for any prediction market.

Returns sentiment score (-1 to 1), actionable recommendation (bullish/bearish/neutral), and AI confidence level.
Goes beyond raw probability — analyzes market psychology, crowd wisdom, and directional bias.
Use for trade signals, contrarian analysis, or augmenting your own prediction models with market sentiment data.

Input parameters:

- `market_id` (string, required): The market UUID or external_id (ticker)

### `get_market_stats` (~68 tokens)

Get aggregate statistics across all prediction markets — totals, categories, sources, and volume.

Returns total market count, active markets, category distribution, source breakdown (Kalshi vs Polymarket), and aggregate trading volume.
Use for market overview, portfolio allocation decisions, or understanding the prediction market landscape.

### `get_trending` (~139 tokens)

Get prediction markets with the biggest probability swings — momentum detection for trending events.

Finds markets where odds moved most in the last 1h, 24h, 7d, or 30d. Surfaces breaking events, sentiment shifts, and market-moving news.
Use when looking for actionable opportunities, volatile markets, or events where consensus is rapidly changing.
Returns markets ranked by absolute probability change with direction (up/down) and current odds.

Input parameters:

- `limit` (number): Maximum markets to return (default: 10, max: 25)
- `timeframe` (string): Lookback window for detecting swings (default: 24h)

### `compare_sources` (~124 tokens)

Compare prediction odds across Kalshi and Polymarket for the same event — find pricing discrepancies.

Searches for markets matching your query on both Kalshi (CFTC-regulated) and Polymarket, then shows side-by-side probabilities.
Use for arbitrage detection, cross-validating predictions, or understanding how regulated vs unregulated markets price the same event.
Returns matched pairs with probability delta and which source is more bullish/bearish.

Input parameters:

- `query` (string, required): Search query to find matching markets across sources (e.g., 'Trump', 'Bitcoin', 'Fed rate')

### `detect_arbitrage` (~251 tokens)

Detect cross-source arbitrage opportunities between Kalshi and Polymarket.

Scans all active markets to find events priced differently across regulated (Kalshi) and unregulated (Polymarket) prediction markets.
Returns actionable opportunities sorted by spread size, with buy/sell signals for each side.

Academic research shows $40M+ extracted from prediction market mispricings annually. Cross-source spreads are structural — different regulation, user bases, and liquidity create persistent pricing gaps.

Use when looking for:
\- Arbitrage opportunities between prediction market exchanges
\- Mispriced markets where consensus disagrees across sources
\- Risk-free profit opportunities from cross-source spread trading

Example: If Kalshi prices "BTC $200K" at 35% and Polymarket at 28%, that's a 7% spread — buy YES on Polymarket, sell YES on Kalshi.

Input parameters:

- `category` (string): Filter by category (default: all)
- `limit` (number): Maximum arbitrage opportunities to return (default: 10, max: 25)
- `min_spread` (number): Minimum probability spread percentage to flag as arbitrage (default: 5, range: 1-50)

### `get_signal` (~214 tokens)

Get a structured pre-computed trading signal for any prediction market — TeleKash Probability Format (TPF).

Combines probability, confidence, sentiment, noise filter, and cross-source data into one actionable signal. This is the complete intelligence package for autonomous agents.

Returns:
\- probability with confidence grade (HIGH/MEDIUM/LOW/VERY_LOW)
\- sentiment score with recommendation (bullish/bearish/neutral)
\- noise filter (signal/weak/noise) — is this momentum real or random walk?
\- cross-source spread (if market exists on multiple exchanges)
\- actionable verdict: STRONG_BUY / BUY / HOLD / SELL / STRONG_SELL / NO_SIGNAL

Use this as the single entry point when an agent needs to make a trade decision. One call replaces get_probability + get_sentiment + get_history + compare_sources.

Input parameters:

- `market_id` (string): The market UUID or external_id (ticker)
- `query` (string): Natural language query to find the market (alternative to market_id)

### `track_prediction` (~168 tokens)

Record a prediction for performance tracking. Agents can log their predictions and later check accuracy via get_performance.

Records: which market, predicted outcome (YES/NO), predicted probability, and confidence level. When the market resolves, your Brier score and calibration are computed automatically.

Use this to build a track record. Agents with verified accuracy get higher trust scores.

Input parameters:

- `agent_id` (string, required): Your agent identifier (any string — use consistently across predictions)
- `market_id` (string, required): The market UUID or external_id
- `predicted_outcome` (string, required): Your predicted outcome
- `predicted_probability` (number, required): Your estimated probability (0.0-1.0) that YES wins. Required for Brier score.
- `reasoning` (string): Brief reasoning for the prediction (optional)

### `get_performance` (~135 tokens)

Get prediction performance metrics for an agent. Shows accuracy, Brier score, calibration, and prediction history.

Returns:
\- Total predictions and resolution rate
\- Accuracy (% correct)
\- Brier score (0 = perfect, 1 = worst — lower is better)
\- Calibration curve (predicted probability vs actual outcome rate)
\- Recent predictions with outcomes

Use this to evaluate an agent's forecasting ability or track your own improvement over time.

Input parameters:

- `agent_id` (string, required): The agent identifier to check performance for
- `limit` (number): Number of recent predictions to return (default: 20, max: 100)

### `get_divergences` (~202 tokens)

Find markets where prediction sources disagree — the highest-value signal in forecasting.

When Kalshi, Polymarket, and Metaculus show different probabilities for the same event, at least one source is wrong. This tool finds those disagreements, ranked by spread size.

Returns:
\- Markets with the largest cross-source probability gaps
\- Which source says what
\- Forecaster count from Metaculus (crowd wisdom depth)
\- Divergence classification: STRONG (>15%), MODERATE (8-15%), WEAK (3-8%)

These are the markets where alpha exists. When sources converge, the edge disappears.

Input parameters:

- `category` (string): Filter by category (crypto, politics, economics, sports, weather, other)
- `limit` (number): Number of divergences to return (default: 10, max: 50)
- `min_spread` (number): Minimum probability spread to include (default: 5 = 5%)

### `get_edge` (~227 tokens)

Capital efficiency analysis — find markets with the best risk/reward for a given bankroll.

Uses Kelly Criterion to compute optimal position sizes and expected value. Returns markets ranked by edge (expected profit per dollar risked).

For each market:
\- Edge = your estimated probability minus market probability
\- Kelly fraction = optimal % of bankroll to allocate
\- Expected value per dollar risked
\- Risk classification (conservative/moderate/aggressive)

Use this when an agent has limited capital and needs to maximize expected returns. Pairs with get_signal for probability estimates and track_prediction for accuracy tracking.

Input parameters:

- `agent_id` (string): Agent ID — uses your prediction history to estimate your edge (optional but recommended)
- `bankroll` (number): Total capital available for allocation (in dollars, default: 1000)
- `category` (string): Filter by category
- `limit` (number): Number of opportunities to return (default: 10, max: 30)
- `min_confidence` (string): Minimum confidence grade to include (HIGH, MEDIUM, LOW — default: MEDIUM)

### `create_market` (~254 tokens)

Create a custom prediction market on TeleKash. Markets are binary YES/NO questions that resolve on a specified date.

Markets created via this tool are tagged as "agent-created" and appear alongside Kalshi/Polymarket/Metaculus markets. Other agents can query, predict on, and trade these markets.

Requirements:
\- Clear YES/NO question in the title
\- Resolution date in the future
\- Category for discoverability
\- Resolution criteria (how to determine the outcome)

Created markets start with 50/50 odds. Probability moves as predictions come in.

Input parameters:

- `category` (string, required): Market category
- `closes_at` (string, required): When trading closes (ISO 8601 datetime)
- `creator_id` (string, required): Agent identifier creating this market (used for attribution)
- `description` (string): Detailed description and context for the market
- `resolution_criteria` (string, required): How the outcome will be determined (e.g., 'Based on CoinGecko BTC price at midnight UTC')
- `resolves_at` (string, required): When the market resolves (ISO 8601 datetime, must be after closes_at)
- `title` (string, required): The prediction question (should be answerable with YES or NO)

## Diagnostics

Captured diagnostic sections: Provenance, Dependencies. The full working is on the page: https://verifymcp.io/servers/telekashoracle-mcp-server/telekash-mcp-server#diagnostics

## Score history

- 2026-08-03: 68
- 2026-08-02: 68
- 2026-08-01: 19
- 2026-07-31: 5
- 2026-07-30: 43
- 2026-07-28: 43
- 2026-07-27: 43

## Links

- npm package: https://www.npmjs.com/package/telekash-mcp-server
- Socket report: https://socket.dev/npm/package/telekash-mcp-server
- Repository: https://github.com/TeleKashOracle/mcp-server
- Changelog RSS feed: https://verifymcp.io/servers/telekashoracle-mcp-server/telekash-mcp-server/changelog.xml
- Changelog JSON feed: https://verifymcp.io/servers/telekashoracle-mcp-server/telekash-mcp-server/changelog.json
- HTML version of this page: https://verifymcp.io/servers/telekashoracle-mcp-server/telekash-mcp-server