d-research-ultra-skill

Prebuilt multi-agent deep research skill for AI agents. Use for source-backed web research, public-data collection, literature reviews, due diligence, policy or standards analysis, market or technical research, evidence ledgers, execution gates, blocker reports, and audit-grade synthesis. Ships the D Research core methodology plus a runtime-neutral orchestrator and six bundled worker-agent role definitions. Uses ephemeral parallel or sequential workers when the host supports them, falls back to manual role checklists when it does not, and creates persistent workers only when the user explicitly requests installation.

Use when editing D Research Ultra helper scripts, references, package metadata, CI, or internal links. Runs the focused checks that protect script syntax, internal references, self-tests, and local artifact hygiene.