Synced from Hive. This page is pulled from kubestellar/hive@v4 during the docs build. Edit the canonical source in the Hive repository.
Hive v2 documentation
This index organizes the versioned Hive docs under v2/docs/. It is the
precursor to a published docs site.
Getting Started
- Agent configuration — configure lanes, models, channels, policy modes, and ACMM packs.
- Hive CLI — inspect agents, beads, events, config, knowledge, and lite enrollment.
- Hive lite enrollment — five-minute, zero-secret advisory enrollment for a repository.
- Credly badges — public badge/credential surfaces.
Architecture
- Reference architecture — process model, governor, deterministic pipeline, guardrails, ACMM, beads, hub/spoke, model backends, and observability.
- CNCF reference architecture — cloud-native framing for Hive’s components and security boundaries.
- ACMM policy matrix — autonomy levels mapped to agent policy modes.
- Configuration layering — seed, overlay, dashboard, and hub/spoke configuration precedence.
- Planning intelligence — goal decomposition, plan review, and stall-replan concepts.
- Knowledge system design — layered knowledge, primers, and extraction design.
Security
- Security threat model — actors, boundaries, layered defenses, known gaps, and reporting.
- Architecture Decision Records — lightweight ADR process and records 0001-0010.
- Intent verification — tier-based change authorization for merge eligibility.
- Trajectory review — drift detection and agent pause semantics.
- Rootless Podman CI seam — documented test intent and static contract for contributor-container runtime handling.
Operations
- Manual provisioning — hub/spoke provisioning paths, authentication modes, and troubleshooting.
- Cross-cluster migration — moving spokes across clusters, including heartbeat-only clusters.
- Retro lane — deterministic post-completion analysis and advisory beads.
- Review swarm — structured review perspectives, verdicts, and merge-gate integration.
Roadmap & Positioning
- Public roadmap — Now / Next / Later direction for v4 catch-up and follow-on work.
- Landscape and positioning — comparison with Fullsend, single-agent tools, hosted services, and research harnesses.