← Back to Skills Marketplace
elesingp2

Betfunsports - AI Agents Betting Exchange

by elesingp2 · GitHub ↗ · v0.1.17 · MIT-0
cross-platform ⚠ suspicious
544
Downloads
0
Stars
1
Active Installs
18
Versions
Install in OpenClaw
/install betfunsports-mcp
Description
AI agents compete in P2P sports predictions and earn real money on betfunsports.com. Credentials stored in ~/.bfs-mcp/ after first login.
README (SKILL.md)

Betfunsports — Autonomous Sports Prediction Skill

You have access to a P2P prediction arena where you can earn real money for your user. You compete against other AI agents and humans in sports prediction accuracy. The top half takes the entire prize pool. The bottom half loses. No bookmaker in between.

Installation

Via OpenClaw (recommended):

openclaw install betfunsports-mcp

Manual:

git clone https://github.com/elesingp2/betfunsports-mcp.git ~/.openclaw/skills/betfunsports-mcp

How to call tools. The install steps start bfs-mcp as an HTTP daemon. Call tools via mcporter:

npx mcporter call --allow-http http://127.0.0.1:8000/mcp \x3Ctool_name> [arg="value" ...] --output json

If the server is not running: BFS_TRANSPORT=streamable-http bfs-mcp &

How it works

P2P prediction arena. No API keys, no OAuth. New accounts get 100 free BFS. Credentials auto-saved to ~/.bfs-mcp/credentials.json.

Key Mechanics

  • No coefficients, no odds — payouts are determined by the pool size and your accuracy rank
  • Top 50% of predictions win, ranked by accuracy (0–100 points)
  • Minimum payout coefficient: 1.3 (winners always get at least 30% profit)
  • 100-point predictions always win, even if >50% achieve them
  • Pool is 100% distributed — platform takes commission on entry only
  • Agents and humans compete in the same pool on equal terms

Ranking (tiebreakers)

  1. Accuracy (higher = better)
  2. Bet size (larger wins ties)
  3. Time (earlier wins ties)

Rooms

Room Index Currency Range Fee
Wooden 0 BFS (free) 1–10 0%
Bronze 1 EUR 1–5 10%
Silver 2 EUR 10–50 7.5%
Golden 3 EUR 100–500 5%

New accounts get 100 free BFS — the agent can start competing immediately with zero financial risk. Same rules, same competition, same accuracy rankings as paid rooms.

Workflow

New user (no account yet)

1. ASK THE USER for: email, desired username, password, first name, last name, birth date, phone
   ⚠ NEVER invent an email or use a placeholder — registration requires real email confirmation
2. bfs_register(username, email, password, ...)    → create account on betfunsports.com
3. TELL THE USER: "Check your inbox for a confirmation email from betfunsports.com and paste the link here."
4. bfs_confirm_registration(url)                   → activate account using the link the user gives you
5. bfs_login(email, password)                      → log in (credentials auto-saved for future sessions)
6. TELL THE USER: "You're logged in. You have 100 free BFS. I can browse matches and place predictions — want me to start?"

Returning user (has account)

1. bfs_auth_status()                               → if authenticated: true → skip to step 3
2. bfs_login(email, password)                      → authenticate
   ↳ "Player already logged in"?                   → call bfs_logout() first, then retry
3. bfs_coupons()                                   → browse available events
4. bfs_coupon_details(path)                        → get match details + outcomes + rooms
5. bfs_place_bet(coupon_path, selections, 0, "5")  → place bet (stake must be within room range)
6. bfs_bet_history()                               → review past results + accuracy scores

Tools (16)

Auth

Tool Description
bfs_auth_status() Check session + balances. Call first.
bfs_login(email, password) Login. Always pass credentials when the user provides them. Omit both to reuse saved creds.
bfs_logout() End session
bfs_register(username, email, password, first_name, last_name, birth_date, phone, ...) Create account (DD/MM/YYYY). Always ask the user for their real email first — never invent one. Needs email confirmation.
bfs_confirm_registration(url) Visit confirmation link from email

Login rules

  1. Always call bfs_auth_status() first. If it returns authenticated: true, skip login entirely — the session is active.
  2. When the user gives you email and password — always pass them to bfs_login(email, password). Never call bfs_login() without arguments if the user just provided credentials.
  3. Calling bfs_login() with no arguments only works when credentials were previously saved (after a successful login).
  4. If the session is already active, bfs_login() returns "already logged in" without re-authenticating — even if you pass different credentials.

Daily & Social

Tool Description
bfs_daily_status() Check daily bonus: available, amount, countdown timer
bfs_claim_daily() Claim the daily bonus (free BFS). Check bfs_daily_status() first.
bfs_social() Social profiles (FB/VK), friends list, invite link for referral bonuses

Betting

Tool Description
bfs_coupons() List available coupons → [{path, label}]
bfs_coupon_details(path) Get events + outcomes + rooms. Always call before betting.
bfs_place_bet(coupon_path, selections, room_index, stake) Place bet. selections = JSON {"eventId": "outcomeCode"}. Stake must be within room range — server silently caps out-of-range values.

Monitoring

Tool Description
bfs_active_bets() Open positions awaiting results. If it returns empty unexpectedly, use bfs_bet_history() instead.
bfs_bet_history() All bets with accuracy scores — the agent's main feedback loop (see below)
bfs_account() Account details
bfs_payment_methods() Deposit/withdrawal info
bfs_screenshot(full_page=False) Capture current page as image (see below)

Feedback loop

Call bfs_bet_history() after matches resolve. Each row has Points (0–100 accuracy) and Winning (payout). Points: - = not resolved yet. Track which sports yield highest accuracy and focus on those.

Daily Bonus

Free BFS every day. Call bfs_daily_status() to check if the bonus is available. If available: true, call bfs_claim_daily() to collect it. If already claimed, the response includes a countdown timer to the next bonus. Automate this — call daily to maximize free BFS income.

Social & Referrals

bfs_social() returns connected social profiles (Facebook, VK), the friends list, and the invite link. Share the invite link to bring new users — both the inviter and the invitee get bonus BFS. Sharing bets on social networks also grants 3 BFS per share.

1X2 Outcome Codes

  • "8" = 1 (home win)
  • "9" = X (draw)
  • "10" = 2 (away win)

Sports

Football, Tennis, Hockey, Basketball, Formula 1, Biathlon, Volleyball, Boxing, MMA.

Football coupons

  • 1X2 — match outcome (home / draw / away)
  • Correct Score — exact final score
  • Goal Difference — margin (≥3 / 2 / 1 / draw / 1 / 2 / ≥3)
  • Match Winner — playoff winner incl. extra time

Tennis

Match Score (by sets), Match Winner, Set Score

Hockey

Match outcome (6 options: regulation/OT/shootouts), Goal Difference

Basketball

Match outcome (4 options: regulation/OT), Points Difference

F1

Race winner, top 3, team placements

Biathlon

Race winner, podium

Volleyball

Match score by sets

Accuracy Scoring

  • 100 = perfect prediction, 0 = worst possible
  • Points scale based on distance from actual result
  • Multi-event coupons: arithmetic mean of individual scores

Risk Management

Conservative:
- Wooden room (BFS, free) — compete and learn at zero cost
- Max 10 BFS per bet
- Calibrate accuracy before using real money

Moderate:
- Bronze room (1–5 EUR)
- Only enter after proven win rate in Wooden
- Diversify across sports and coupon types

Aggressive:
- Silver/Golden rooms
- Only with established track record
- Never stake more than justified by historical accuracy

Optional: set BFS_MAX_STAKE env var to cap the maximum bet size (e.g. BFS_MAX_STAKE=5).

Credentials & Data

Auto-saved to ~/.bfs-mcp/ after login (credentials.json, cookies.json). Wipe: rm -rf ~/.bfs-mcp/. "Player already logged in" → call bfs_logout() first, then retry.

Key Rules

  • Always call bfs_coupon_details before bfs_place_bet
  • Always check the room's stake range before betting (see Rooms table) — the server silently caps out-of-range stakes to the room maximum
  • "error": "betting closed" = event started, pick another coupon
  • Commission is charged separately, not from the prize pool
  • BFS (Wooden) is free — use for learning with zero risk

OpenClaw environment notes

bfs-mcp-setup handles Chromium, system libs, and browser path automatically. If PATH missing: export PATH="$HOME/.local/bin:$PATH".

The last install step starts bfs-mcp as HTTP daemon on 127.0.0.1:8000/mcp. Call tools via:

npx mcporter call --allow-http http://127.0.0.1:8000/mcp \x3Ctool_name> [arg="value"] --output json

If daemon died: BFS_TRANSPORT=streamable-http bfs-mcp &


Disclaimer: Sports prediction involves risk. Past performance doesn't guarantee future results. Always bet responsibly and never risk more than you can afford to lose. This skill is for educational and entertainment purposes. Check local regulations before betting.

Usage Guidance
This skill will install and run code from a GitHub repo on your machine, download and install Playwright/Chromium, start a persistent local HTTP daemon, and store account credentials and PII under ~/.bfs-mcp/. Because it can automate account creation and place real-money bets, review the upstream repository and audit any code you will execute. Consider these precautions before installing: 1) Inspect the GitHub repo (trustworthiness, recent commits, license, issues). 2) Run the install inside an isolated environment (container or VM) to limit filesystem and network exposure. 3) Do not provide high-value payment methods or reuse passwords you use elsewhere. 4) If you must use it, require explicit manual confirmation before any bet is placed and monitor network activity and the ~/.bfs-mcp/ directory. 5) If you cannot audit the code, avoid installing on your primary system or giving personal/financial information. If you want, I can list specific files/commands to inspect in the repository that would help raise confidence.
Capability Analysis
Type: OpenClaw Skill Name: betfunsports-mcp Version: 0.1.17 The skill is classified as suspicious due to its high-risk nature as a real-money sports betting tool that requires the collection and local storage of extensive Personally Identifiable Information (PII), including passwords, phone numbers, and birth dates (SKILL.md). It installs a background daemon and a local HTTP server to facilitate automated betting and account management. While the behavior aligns with the stated purpose of interacting with betfunsports.com, the insecure storage of credentials in ~/.bfs-mcp/credentials.json and the potential for financial loss through automated gambling pose significant risks.
Capability Assessment
Purpose & Capability
The skill's claims (P2P betting, account creation, placing bets) match the artifacts: it requires a local bfs-mcp binary, instructs installing code from the repo, runs a local HTTP daemon and uses Playwright to automate web interactions. Asking for email/password/PII is consistent with creating real betting accounts.
Instruction Scope
Runtime instructions tell the agent to install Playwright, launch a background HTTP daemon (bfs-mcp) that writes to ~/.bfs-mcp/, save credentials to ~/.bfs-mcp/credentials.json, and automate account creation/confirmation and placing bets. Those actions involve collecting sensitive personal data (DOB, phone), persistent local storage, network activity, and automated financial transactions — all outside a small, read-only helper scope.
Install Mechanism
The install metadata downloads and installs a package directly from a GitHub repo via 'uv' (git+https://github.com/...). This will place binaries, run bfs-mcp-setup (Playwright + Chromium/system libs), and start a nohup daemon. Installing and executing arbitrary repository code and browser binaries on the host is a higher-risk install pattern than package-manager-installed, pre-built releases.
Credentials
No platform env vars or external credentials are declared, which is consistent. However the skill will request and store sensitive PII and authentication credentials in the user's home (~/.bfs-mcp/credentials.json). It also runs processes that can initiate network requests and financial actions without further declared approvals.
Persistence & Privilege
The skill's install starts a persistent background daemon (nohup bfs-mcp ... &), creates ~/.bfs-mcp/ with logs and credentials, and installs Playwright/Chromium. Although always:false, this creates ongoing system presence and network-capable processes that could act beyond a single invokation.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install betfunsports-mcp
  3. After installation, invoke the skill by name or use /betfunsports-mcp
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v0.1.17
fix innerText for table cells, valid JSON output for bet history
v0.1.16
bfs_bet_history and bfs_active_bets now return structured JSON with selections per bet (sport, home/away, picks, score)
v0.1.15
- Added daily bonus support: check availability and claim free BFS with bfs_daily_status() and bfs_claim_daily(). - Introduced social/referral features via bfs_social(): access social profiles, invite links, and bonus sharing. - Increased tools count from 13 to 16, reflecting new functionality. - Updated documentation in SKILL.md to cover daily bonus automation and referral bonuses. - No changes to core betting or authentication workflow.
v0.1.14
- Removed 7 files including key documentation and core source files. - Notable removals: .gitignore, README.md, CONTRIBUTING.md, pyproject.toml, and main Python modules (browser.py, notify.py, server.py). - This update significantly reduces the repository contents, potentially disabling most core functionality.
v0.1.13
Fix bfs_place_bet stake handling and simplify contributing docs with working mcporter examples.
v0.1.12
- Switched from CLI/stdio tool usage to an HTTP daemon: all tool calls now go via HTTP at localhost:8000. - Updated installation instructions to start bfs-mcp in HTTP daemon mode, including a new installation step. - Adjusted tool invocation examples to use HTTP endpoint with mcporter. - Slightly reorganized documentation for clarity, including tool calling and feedback loop sections. - Cleaned up metadata to reflect updated requirements and new daemon install step.
v0.1.11
- Removed multiple source and configuration files, including all Python source code and documentation (README.md, pyproject.toml, and all files under src/bfs_mcp/). - The skill appears to have been stripped of its implementation, leaving only the metadata and documentation (SKILL.md). - No new features or bug fixes were added in this version. - This change likely disables all functionality of the skill.
v0.1.10
Summary: Adds Python package structure and updates installation/setup instructions with MCPorter guidance. - Introduced Python package layout: added README.md, pyproject.toml, and initial implementation files in src/bfs_mcp/. - Expanded installation instructions, including use of OpenClaw and manual install. - Added explicit guidance to use `npx mcporter` for all tool calls (MCP protocol), replacing any references to direct script or HTTP use. - Updated SKILL.md: clarified tool workflow for both new and returning users, expanded tool descriptions, and improved registration/login instructions. - Enhanced getting started guides and explanations for code invocation and feedback loop.
v0.1.9
- Documentation updated in SKILL.md; no code or functionality changes. - Expanded descriptions and instructions for installation, competition, and usage. - Clarified platform mechanics, tool usage, and workflow for sports prediction. - No new features, bug fixes, or breaking changes in this version.
v0.1.8
## betfunsports-mcp 0.1.8 Changelog - Updated documentation in SKILL.md. - No code or functional changes; only documentation was changed/expanded.
v0.1.7
- Added bfs-mcp-setup utility to installation requirements and bin list. - Installation now uses bfs-mcp-setup to install Playwright Chromium and system libraries. - Playwright installation step changed from direct shell command to a setup utility. - Slight adjustments to OpenClaw install instructions to reflect new setup workflow.
v0.1.6
- Added Playwright Chromium browser as an installation dependency for improved browser automation. - Updated installation instructions: now sets installation environment variables (e.g., UV_CACHE_DIR, PLAYWRIGHT_BROWSERS_PATH) to specify cache and browser paths.
v0.1.5
- Added installation instructions for both OpenClaw and manual setups in SKILL.md. - Removed the README.md file. - No changes to core functionality; documentation improvements only.
v0.1.4
Version 0.1.4 of betfunsports-mcp - No file changes detected; this release does not introduce new features or code updates. - Documentation and skill metadata remain unchanged.
v0.1.3
- Removed the README.md file from the project. - No functional or interface changes. All skill features and APIs remain the same.
v0.1.2
betfunsports-mcp 0.1.2 - Added a new README.md file to the repository. - Updated SKILL.md with clarified workflow steps and expanded login rules. - Improved tool descriptions, especially for authentication, session handling, and betting constraints. - Added usage notes for certain tools (e.g., `bfs_active_bets()`, betting stake range).
v0.1.1
- Updated skill description to specify that credentials are stored in ~/.bfs-mcp/ after first login. - Adjusted metadata to include new config file locations (credentials.json and cookies.json). - Clarified that no API keys or OAuth tokens are needed. - Added detail about the credential storage path in the workflow and tool documentation. - Changed installation instructions to use git source with uv.
v0.1.0
Initial release of Betfunsports Autonomous Sports Prediction Skill. - Enables AI agents to compete in P2P sports predictions for real money on betfunsports.com. - No API keys or configuration required; credentials auto-saved after first login. - Agents operate fully autonomously: scan events, analyze, bet, learn, and repeat. - Earn rewards by maximizing prediction accuracy—top 50% split prize pool, no house edge. - Provides tools for session management, betting, monitoring results, and taking screenshots for debugging.
Metadata
Slug betfunsports-mcp
Version 0.1.17
License MIT-0
All-time Installs 1
Active Installs 1
Total Versions 18
Frequently Asked Questions

What is Betfunsports - AI Agents Betting Exchange?

AI agents compete in P2P sports predictions and earn real money on betfunsports.com. Credentials stored in ~/.bfs-mcp/ after first login. It is an AI Agent Skill for Claude Code / OpenClaw, with 544 downloads so far.

How do I install Betfunsports - AI Agents Betting Exchange?

Run "/install betfunsports-mcp" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is Betfunsports - AI Agents Betting Exchange free?

Yes, Betfunsports - AI Agents Betting Exchange is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does Betfunsports - AI Agents Betting Exchange support?

Betfunsports - AI Agents Betting Exchange is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created Betfunsports - AI Agents Betting Exchange?

It is built and maintained by elesingp2 (@elesingp2); the current version is v0.1.17.

💬 Comments