Openclaw Skill
WarnAudited by ClawScan on May 10, 2026.
Overview
The skill appears to match its stated purpose, but it can place real Polymarket trades using your prob.trade credentials without built-in confirmation or spending limits.
Review carefully before installing. This is not evidence of malware, but it is a real trading integration: only provide prob.trade credentials if you are comfortable with the agent being able to read balances/positions and place or cancel orders. Keep keys revocable, protect the config file, and require manual confirmation for every trade.
Findings (2)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
If the agent misunderstands a request or is induced to trade, it could place real orders and put the user's funds at risk.
The order command immediately submits an authenticated trading request after constructing the order body, with no local confirmation step, spending cap, or price/amount guardrail shown in the artifacts.
data = trading_request("POST", "/order", body)Use this skill only with explicit user approval before every trade, prefer low-limit or revocable API keys, and consider adding local confirmation and maximum notional limits before enabling trading.
Anyone or any agent process with access to these credentials may be able to view account information and submit trades through prob.trade.
The client reads a prob.trade API key and secret from the environment or config file and uses them to sign requests; this is expected for the service, but it grants delegated access to account data and trading actions.
api_key = os.environ.get("PROBTRADE_API_KEY", "")
api_secret = os.environ.get("PROBTRADE_API_SECRET", "")Store the config file securely, use revocable or scoped keys if prob.trade supports them, and revoke the key from the dashboard if you no longer use the skill.