The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides a configuration script (agents/setup.sh) that manages the placement of agent definition files using standard shell commands such as awk and mkdir. It also provides examples for custom gateway scripts using curl and jq to facilitate communication with AI model APIs.\n- [EXTERNAL_DOWNLOADS]: The orchestration workflow facilitates communication with external AI providers and gateways, including Anthropic, OpenAI, Google, xAI, OpenRouter, and Vercel AI Gateway. These interactions are a core requirement for the skill's functionality as a multi-model coordinator.\n- [PROMPT_INJECTION]: The skill represents an indirect injection surface as it is designed to ingest and process data from external sources, including web content (via the Researcher agent) and existing codebases. It implements a defensive architecture by including a specialized 'Adversarial Reviewer' agent role, which is explicitly tasked with identifying security flaws and malicious logic in the output of other agents.

agent-collaboration