← Back to Skills Marketplace
Prediction Market Watcher
by
m-lwatcher
· GitHub ↗
· v1.0.0
· MIT-0
81
Downloads
0
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install prediction-market-watcher
Description
Monitor, analyze, and trade on Kalshi and Polymarket prediction markets. Use when the user wants to check open positions, scan for value bets, place bets, ch...
Usage Guidance
Do not install or run this skill with live credentials until you are comfortable with the code and where secrets will be stored. Specific actions to consider:
- The registry metadata does NOT list required credentials, but the code and SKILL.md require KALSHI_KEY_ID and a private key file path (KALSHI_KEY_FILE or kalshi-agent/config.json). Treat that as a red flag: ask the publisher to fix the metadata or update it yourself.
- The skill will read an RSA private key file from your workspace and persist trading state (risk_state.json). Store keys in a secure location (not in a shared repo), restrict file permissions, or use a secure secrets store. Do not commit keys to source control.
- Run first in demo mode (the code supports demo=true) and inspect network traffic (which should go only to the Kalshi API base URLs present in the code) before granting live credentials.
- Audit the code if you plan to enable auto-betting (--run). Pay attention to create_order flows and error handling. Consider throttling/limiting auto-bet behavior and keep tight daily caps.
- If you need external notifications (Telegram) or price sources (Brave API), implement and audit them explicitly rather than following the informal SKILL.md instructions.
- If you are not comfortable reviewing code: do not supply live API keys or private keys. Ask the skill publisher to correct the metadata and clarify how secrets are stored and used.
Capability Analysis
Type: OpenClaw Skill
Name: prediction-market-watcher
Version: 1.0.0
The skill bundle is a functional trading agent for Kalshi and Polymarket prediction markets. It includes a robust API client (kalshi_client.py) with RSA-PSS signing, a market analysis engine (analyzer.py), and a dedicated risk management module (risk.py) that enforces strict betting limits ($20/day, $5/bet) and position caps. No evidence of data exfiltration, malicious execution, or harmful prompt injection was found; the code is well-structured and aligns entirely with its stated purpose of assisting a user in monitoring and placing bets on regulated exchanges.
Capability Assessment
Purpose & Capability
The skill claims to monitor and trade on Kalshi/Polymarket and the code implements that. However the registry metadata lists no required environment variables or credentials while both SKILL.md and the code clearly require KALSHI_KEY_ID and KALSHI_KEY_FILE (an RSA private key file) or a config with key_id/key_file. That omission is an incoherence: a trading skill legitimately needs credentials, so the registry should declare them. Also SKILL.md refers to storing keys at kalshi-agent/kalshi.key in the workspace; requesting a private key file is proportionate for the function but must be declared.
Instruction Scope
SKILL.md and the scripts are scoped to market scanning, ranking, and order placement. They instruct reading a workspace config and private key, fetching market data, and writing a local risk_state.json. The doc also suggests external actions (cron job, Telegram messages, Brave API/web search for prices) that are outside the provided code. The provided code itself only talks to Kalshi endpoints and reads/writes local state; it does not implement Telegram or Brave calls. The instructions asking the operator to message an external person (Katie) are out-of-band and not enforced by code — this is a procedural note but not inherently malicious.
Install Mechanism
There is no install spec — the skill is instruction+scripts only. That is lower install risk (no downloads from arbitrary URLs). The package contains Python scripts that will run from the workspace; nothing in the manifest downloads or extracts remote archives.
Credentials
The runtime requires sensitive artifacts (Kalshi API key ID and an RSA private key file) and will read them from env vars or a workspace config, but the registry metadata did not declare these required env vars/credentials. The code also persists trading state and open positions to risk_state.json in the script directory. Requesting Kalshi credentials is proportionate to the trading purpose, but the missing declaration plus the need to place a private key file in the workspace increases the chance a user will accidentally store secrets in an insecure place. No unrelated credentials are requested by code.
Persistence & Privilege
The skill writes and reads local persistent state (risk_state.json) and expects/stores a config.json and private key file in the workspace. It does not request 'always: true' and does not alter other skills or system-wide settings. Local persistence and requiring a private key are normal for a trading bot, but you should be aware these files contain sensitive info and trading history and will live in the workspace by default.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install prediction-market-watcher - After installation, invoke the skill by name or use
/prediction-market-watcher - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Initial release — Kalshi portfolio tracking, market scanning, bet placement, settlement reminders. Polymarket read-only reference included.
Metadata
Frequently Asked Questions
What is Prediction Market Watcher?
Monitor, analyze, and trade on Kalshi and Polymarket prediction markets. Use when the user wants to check open positions, scan for value bets, place bets, ch... It is an AI Agent Skill for Claude Code / OpenClaw, with 81 downloads so far.
How do I install Prediction Market Watcher?
Run "/install prediction-market-watcher" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Prediction Market Watcher free?
Yes, Prediction Market Watcher is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Prediction Market Watcher support?
Prediction Market Watcher is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Prediction Market Watcher?
It is built and maintained by m-lwatcher (@m-lwatcher); the current version is v1.0.0.
More Skills