Skip to content

VolcanoEngineering/Bamboo

Repository files navigation

Bamboo: The AI Governance OS πŸŽ‹

High-Velocity Governance for Multi-Agent Squads and Orchestration.

A governing AI memory scaffold to catch drift before tech debt becomes a thing and AI memory collapse occurs.


[STATUS] System Pulse: ACTIVE πŸ“‘

Component Status Mandate
Policy STABLE Operational Governance
Watchdog VIGILANT Memory Integrity & Drift Detection
State-Bus PERSISTENT Session Continuity & Recovery
The Canon LOCKED Structural Verification (PLTRF)

The Ethos

Bamboo transforms a standard repository into a Governed Workspace. It provides the stable policy, memory, and resource limits required to manage high-fidelity projects.

We prioritize Physics over Liturgy: if a rule cannot be reduced to a file path, a script exit code, or a resource ceiling, it doesn't belong in the canon.


🏎️ Fork in 5 Minutes

The fastest path to a governed repo is the Use this template button.

1. Establish the Core (Discipline Only)

Drop these into your project root to establish the behavioral baseline:

  • AGENT.md: The cold-start router (MUST include Session Identity).
  • Bamboo.md: The operational policy layer.
  • behavior/: The 3-concept ruleset and decoder ring.
  • docs/ctx-orientation.md: Your project's narrative history.

The core discipline is vendor-neutral and dependency-free.

2. Establish the OS (Full Governance)

For multi-agent or complex environments, the Bamboo OS runtime extension provides the high-fidelity chassis for lifecycle management, integrity heartbeat, and event-driven sync.

Access to the runtime engine is managed via the private BAMBOO-OS repository.


πŸ—οΈ Core Architecture (The 3-Concept Canon)

To protect cold-start economics, Bamboo mandates only three load-bearing concepts. Everything else is an extension.

  1. Knobs ⭐: The unit of narrative state change. Every meaningful architectural state transition earns a dated, narrative entry in the orientation log. We log the "Why", not just the code.
  2. PLTRF πŸ—οΈ: Preventative Long-Term Repo Fragmentation. The binary discipline of structural integrity. One canonical home per concept. Broken references, decaying links, or orphaned files are build-failing compilation errors.
  3. Tiering ❄️: Memory preservation via Hot/Warm/Cold/Ice classification. Keep working memory lean, archive the rest, and freeze files > 6 months old to prevent retrieval drift.

πŸ›‘οΈ Strategic Extensions (Optional)

  • architecture/: Advanced memory systems (ADM/RAG/Watchdog/OKI).
  • BAMBOO-OS: The private runtime engine (Scripts, Schemas, Multi-agent Topology).
  • skills/: Portable AI capabilities and vendor-specific overlays.
  • design/: UI/UX standards and component rules.

πŸ“š Intellectual Grounding

Bamboo is a synthesis of established work across information theory, cognitive science, and AI engineering:

  • Information entropy β€” Shannon (1948)ΒΉ. Entropy as a measure of disorder gives the vocabulary for context entropy.
  • Working memory limits β€” Miller (1956)Β². The "magical number seven" establishes the need for Knob-based chunking.
  • Cognitive load theory β€” Sweller (1988)Β³. PLTRF discipline is a sustained attack on extraneous cognitive load.
  • Retrieval-Augmented Generation (RAG) β€” Lewis et al. (2020)⁴. Bamboo is a structural RAG complement.
  • Long-context degradation β€” Liu et al. (2023)⁡, "Lost in the Middle." Empirical validation for the "newest at the top" Knob log rule.

For the deep theoretical treatment, see Documentation.md.


πŸ† Recent Milestones

  • 2026-07-01: Formally shakedown-tested the Operational Knowledge Integrity (OKI) and Repository Understanding Health (RUH) frameworks in multi-agent simulation and orchestration environments.
  • 2026-06-13: Bamboo has a CI gate that fails the build when an agent violates a mandatory rule. Partitioned core discipline and OS into separate repositories. Finalized Governance Core and Session Identity Clause.
  • 2026-06-03: Introduced root Bamboo.md and separated policy from routing.
  • 2026-05-28: Added Memory Watchdog and Drift to the architecture layer.

Created and maintained by Matt Stenquist. Special thanks to testing contributors for production-testing the chassis.

References

  1. Shannon, C. E. (1948). "A Mathematical Theory of Communication." Bell System Technical Journal.
  2. Miller, G. A. (1956). "The Magical Number Seven, Plus or Minus Two." Psychological Review.
  3. Sweller, J. (1988). "Cognitive Load During Problem Solving: Effects on Learning." Cognitive Science.
  4. Lewis, P. et al. (2020). "Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks." NeurIPS.
  5. Liu, N. F. et al. (2023). "Lost in the Middle: How Language Models Use Long Contexts." arXiv:2307.03172.

License

Bamboo is source-available under the PolyForm Noncommercial License 1.0.0 β€” free for noncommercial use (personal, research, evaluation, education, nonprofit). Any commercial use requires a separate commercial license; see COMMERCIAL.md.

About

Your Multi-Agent Orchestrator for AI Governance in AI-assisted repos, and projects. πŸ—οΈπŸ§±

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages