socket-fix

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt embeds a concrete API token value (the demo token) and shows commands that set it directly (npx socket config set apiToken ...), which requires the agent to include a secret verbatim in generated commands/output.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly instructs running a remote install script at runtime via curl -fsSL https://raw.githubusercontent.com/SocketDev/socket-patch/main/install.sh | sh, which fetches and executes remote code and is presented as an installation method required for the patch workflow.

HIGH W008: Secret detected in skill content (API keys, tokens, passwords).

• Secret detected (high risk: 1.00). I scanned the entire skill prompt for literal credentials. The CLI setup section contains a high-entropy, literal API token in the example command:

npx socket config set apiToken sktsec_t_--RAN5U4ivauy4w37-6aoKyYPDt5ZbaT5JBVMqiwKo_api --no-banner --no-spinner

This is not a placeholder (it is not "YOUR_API_KEY" or "sk-xxxx") and is a real-looking token string (high entropy, not truncated). Although the text labels it a "public demo token" with limited permissions, it is nonetheless an actual credential embedded in the documentation. No other high-entropy secrets (API keys, PEM blocks, or confidential tokens) are present; other items are environment variable names or placeholders, which I ignored per the rules.