shodan-reconnaissance

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt explicitly shows and instructs embedding API keys and secret tokens verbatim in CLI commands, curl URLs, and code (e.g., shodan init YOUR_API_KEY, curl ...?key=YOUR_KEY, API_KEY = 'YOUR_API_KEY'), which requires the LLM to handle/output secrets directly and creates an exfiltration risk.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.75). Shodan CLI/REST API runtime calls (e.g., shodan host, shodan search, shodan download, and curl https://api.shodan.io/...) ingest Shodan’s returned host/service/banner text and metadata from a public third-party source (Shodan), which is outsider-authored free-form content that can be placed into the agent’s LLM context.