The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill fetches Freshworks Crayons (UI components) from reputable CDNs (jsdelivr.net and unpkg.com) and provides instructions to install the Freshworks CLI (FDK) from the vendor's official domain (freshdev.io). These are standard operations for building Freshworks applications.
[COMMAND_EXECUTION]: Utilizes the shell tool to run 'fdk validate', 'node --version', and 'fdk version'. These commands are exclusively used to ensure that the user's application environment is correct and that the generated code passes platform-standard validation tests.
[PROMPT_INJECTION]: The skill uses strict instructional language (e.g., 'ZERO TOLERANCE', 'MANDATORY ENFORCEMENT') and role-play personas ('Freshworks Platform 3.0 senior solutions architect'). These are utilized as quality guardrails to ensure the AI agent outputs high-quality, compliant code rather than attempting to bypass safety protocols.
[DATA_EXPOSURE]: The skill includes robust security rules (rules/security.mdc) and a dedicated security auditor agent (agents/app-security-scanner.md) designed to prevent the exposure of API keys, tokens, or PII in the code generated by the agent.

app-dev