NOFX AI Trading

NOFX AI Trading OS integration - crypto market data, AI trading signals, strategy management, trader control, and automated reporting. Use when working with...

What to consider before installing: - This package expects and uses an API key (NOFX_API_KEY) and a config file (default: ~/clawd/skills/nofx/config.json) but the registry metadata does not declare those requirements — do not assume keys are optional. Provide only least-privilege exchange keys (disable withdrawals, restrict permissions and IP whitelist). - The skill’s browser automation expects a profile named 'clawd' (logged-in session). Granting the agent/browser access to a profile can expose other site sessions/cookies; only use a dedicated, isolated browser profile and don't reuse your personal browser profile. - The scripts call external endpoints at nofxos.ai and include examples to post to Telegram/Discord/Slack. Confirm you trust those endpoints and any webhook URLs you configure. - The docs show running remote install scripts via curl|bash from GitHub raw — review those scripts before executing; avoid blind 'curl | bash'. - The shipped shell scripts depend on curl and jq but the metadata does not list required binaries; ensure you inspect and run scripts locally in an isolated environment before allowing the agent to run them. - If you proceed: review config.json and scripts manually, store API keys in a secure secret store (or environment variables with least privilege), use test/demo exchange keys first, and be prepared to rotate/revoke keys if anything suspicious occurs. - If you want a safer assessment, provide the install.sh referenced by the docs or confirm whether the agent will be given access to your browser profile or filesystem; that information would change the risk assessment.

Type: OpenClaw Skill Name: nofx Version: 1.1.0 The skill is classified as suspicious due to significant vulnerabilities, specifically the risk of shell injection and API key exposure. The `scripts/nofx-api.sh` script directly embeds unsanitized arguments (e.g., `symbol`, `limit`, `duration`) into `curl` commands, creating a potential remote code execution (RCE) vulnerability if an attacker can control these inputs. Additionally, the script passes the API key as a URL query parameter (`?auth=$API_KEY`), which is less secure than using an Authorization header and increases the risk of the key being logged or exposed. While the skill's stated purpose is legitimate and there's no evidence of intentional malice, these critical vulnerabilities warrant a 'suspicious' classification.