The Agent Skills Directory

[COMMAND_EXECUTION]: The skill invokes the bdata CLI utility for web discovery and authentication. These actions are consistent with the skill's purpose and use the vendor's specialized tooling to interact with their API.
[COMMAND_EXECUTION]: The skill executes a local script scripts/merge_corpus.sh which uses the jq utility to aggregate JSON results. This script includes logic to deduplicate URLs and filter out low-quality content or automated block pages, which enhances the reliability of the research process.
[SAFE]: While the skill ingests content from external websites (an indirect prompt injection surface), it mandates the use of quality gates, primary source prioritization, and inline citations. These instructions guide the agent to maintain factual grounding and transparency, effectively mitigating the risks associated with processing untrusted web content.

live-research