The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the rg (ripgrep) utility to search the local workspace for existing material related to the research topic. This is a standard and safe use of local tools to provide context before proceeding to external searches.
[EXTERNAL_DOWNLOADS]: The skill orchestrates multiple web searches and content fetches through standard agent tools to aggregate information from diverse sources. This is the primary function of the research pipeline and utilizes expected network capabilities.
[SAFE]: The skill contains extensive safety guidelines for the agent, including mandatory reflection steps (Phase 4), strict verification of all cited sources (Phase 6), and instructions to explicitly surface conflicting information rather than resolving it through speculation. It also includes specific defenses against LLM failure modes like sycophancy and source laundering.

deep-research