The Agent Skills Directory

[SAFE]: The skill defines structural patterns and logical flows for generating CPQ application configurations. It focuses on user interface components, business rule mapping, and role-based access control within a specific framework. No malicious instructions, unauthorized network operations, or sensitive data exposure patterns were detected.
[INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of external domain-specific data (DOMAIN.md) to drive the generation of application logic and user interfaces. This is an inherent part of the skill's purpose to build customized tools. 1. Ingestion points: The agent reads the DOMAIN.md file to extract entities, products, pricing logic, and stakeholder maps. 2. Boundary markers: The skill provides deterministic mapping rules and a strict TypeScript schema (references/config-schema.md) to constrain how the agent interprets and structures the ingested data. 3. Capability inventory: The skill's primary capability is generating a JSON configuration file for a single-page application (SPA). It does not perform arbitrary shell execution or make network requests based on the ingested content. 4. Sanitization: The skill relies on the agent's mapping process to translate natural language into structured configuration data, effectively sanitizing the input against the required application schema.

cpq-builder