geepers-orchestrate
par lukeslp
Web & Frontend DevelopmentRésumé
Run multi-agent Dream Cascade (hierarchical 3-tier synthesis) or Dream Swarm (parallel multi-domain search)
Description
Ce que fait ce skill
Runs multi-agent research workflows via the dr.eamer.dev API. Supports two modes: Dream Swarm dispatches parallel agents across multiple data sources simultaneously, and Dream Cascade runs a three-tier hierarchical pipeline where agents gather, synthesize, and summarize in stages.
Parallelizing across specialized agents compresses what would be many sequential queries into a single API call with a unified synthesized result.
Quand l'utiliser
- Researching a medical topic across PubMed and academic sources
- Synthesizing competitive landscape analysis from multiple domains
- Summarizing the state of a technical field with an executive briefing
- Cross-domain policy research requiring perspectives from multiple disciplines
- Deep-dive analysis on a scientific question that would take many sequential queries
Exemple de workflow
User asks: What are the most effective treatments for Type 2 diabetes?
- 1
Send POST request to /v1/orchestrate/swarm with the query and sources [pubmed, semantic_scholar, wikipedia]
- 2
API spawns 5 parallel agents, each querying a subset of the specified sources
- 3
Agents return domain-specific findings to the orchestrator
- 4
Orchestrator synthesizes all findings into a unified response
A single synthesized answer with source attribution and metadata including agent count and duration
Prérequis
- DREAMER_API_KEY environment variable from dr.eamer.dev
Installation
clawhub install lukeslp/geepers-orchestrateExécute la commande CLI, ou colle l'instruction dans Claude Code, Cursor, Codex ou l'agent de code que tu utilises.