The Agent Skills Directory

[SAFE]: The skill provides detailed instructions for implementing referral programs using trusted third-party services like ReferralCandy and Referral Hero. For custom implementations, it provides TypeScript code that adheres to security best practices.\n- [COMMAND_EXECUTION]: No suspicious or unauthorized command executions were found. The code snippets focus on standard web application logic (middleware, database operations, and reward calculation).\n- [DATA_EXFILTRATION]: No data exfiltration patterns were identified. The network-adjacent operations are limited to standard cookie management and database interactions within a typical e-commerce context.\n- [PROMPT_INJECTION]: The instructions do not contain any patterns designed to override agent behavior, bypass safety filters, or extract system prompts.\n- [OBFUSCATION]: No obfuscated code, hidden characters, or encoded URLs were found. All provided code and configuration files are in plain, readable text.\n- [INDIRECT_PROMPT_INJECTION]: While the skill involves processing external data (referral codes from URL parameters), the implementation logic includes necessary safeguards such as first-click attribution windows, rate limiting at the click-logging layer, and validation of referral eligibility to prevent common abuse vectors.

referral-program