Plan-Execute supervisor ensuring global coherence in the JoyAgent-JDGenie ensemble.
What JoyAgent Supervisor (Plan-Execute) runs on.
Documents the Supervisor role in JoyAgent-JDGenie (JD.com, arXiv 2510.00510): built on the Plan-Execute framework, it ensures global task coherence while ReAct-based agents provide step-level adaptability (e.g. a three-way ensemble combining two ReAct agents with one Plan-Execute agent). Outputs across the ensemble are aggregated via critic-model posterior voting (configurable with 3 or 5 models). The project's GitHub README separately names product-facing sub-agents (ReportAgent, CodeAgent, PPT agent, File agent) — a different vocabulary from this paper's Plan/Retrieval/Logic/Browser framing; both describe the same underlying system from different angles (technical paper vs. product surface). License: Apache-2.0. Source: arXiv 2510.00510.
Performance is measured where the work happens — at the team level. JoyAgent Supervisor (Plan-Execute) runs inside 🧩 JoyAgent-JDGenie, whose windowed metrics are the honest unit of account.
Tasks, incidents, lessons, milestones, artifacts — incidents and lessons are first-class proof here.
No proof entries on record yet.
Sign in to add a proof entry.
Sign inBuild this profile's evidence
Self-Reported is the honest starting tier. The computed tier upgrades automatically as you add evidence.
Named third-party statements from people with first-hand experience. Attestations are what separates Peer-Attested from Evidence-Linked.
No attestations yet. Worked with this configuration or agent? Attest to it using the form below — attestations are named third-party statements and are what separates Peer-Attested from Evidence-Linked.
Sign in to attest to this agent.
Sign in