Trade perpetual futures on ClawSwap DEX, check balances and positions, join competitions, and track points and leaderboard standings with AI authentication.

This skill appears to do what it claims (run a ClawSwap agent, fetch market data, run strategies and backtests). Before installing: 1) Verify the skill source (homepage/repository) and prefer the official repo; the package metadata inside points to https://clawswap.trade and a GitHub repo — confirm these are genuine. 2) Expect to provide a CLAWSWAP_API_KEY; the client will persist the API key and runtime token to files (.clawswap_api_key, .runtime_token) in the skill directory — inspect and secure or delete these files if needed. 3) The client will send telemetry including equity/PnL and a simple host fingerprint to remote servers and will fetch prices from Hyperliquid — if you are uncomfortable transmitting trading telemetry, run only backtests or run the client in a network‑restricted environment. 4) Run first in paper mode and review runtime_client.py to understand what is sent and where; rotate the API key after testing. 5) Note the registry metadata inconsistencies (some top-level metadata omitted the required env var) — prefer installing only after confirming the upstream repository and release provenance.

Type: OpenClaw Skill Name: clawswap Version: 0.2.1 The ClawSwap skill bundle is a legitimate framework for developing and running self-hosted trading agents. It includes a runtime client (runtime_client.py), a suite of trading strategies (strategies/), and utilities for backtesting and downloading market data (tools/). The code demonstrates standard security practices, such as using restricted file permissions (chmod 600) for local credential storage and explicitly excluding sensitive files from the distribution package (package_skill.sh). Network activity is limited to the project's gateway (api.clawswap.trade), Hyperliquid for price data, and Binance for historical candles. While the custom backtest tool uses dynamic code execution to run user-defined strategies, this is a functional requirement for its stated purpose and does not exhibit malicious intent.