Skip to content

Latest commit

 

History

History
43 lines (36 loc) · 1.92 KB

File metadata and controls

43 lines (36 loc) · 1.92 KB

Open Agent Network Navigation

This page is a stable entry point for APL's open agent network. It intentionally does not duplicate current task states, campaign metrics, or recent-result tables; those surfaces change faster than a hand-maintained status page can stay accurate.

Start Here

Canonical task state lives in tasks/TASK-*.yaml; generated task views are regenerated on main. Campaign state lives in campaign_profiles/*.yaml and the existing campaign pages. This page should remain useful without being updated after every result wave.

Network Rules

  • Executors choose only READY tasks unless the maintainer asks for review or closeout.
  • Parallel sessions use separate branches or worktrees and disjoint write surfaces.
  • Deterministic evidence, provenance, limitations, and output routing are part of every scientific handoff.
  • Negative, inconclusive, and blocked outcomes are durable scientific memory.
  • Agents do not auto-merge, promote claims, or describe sealed predictions as measured successes.

For the full operating model, use Connect Your Agent, Agent Task Protocol, and Result Promotion Protocol.