tony/bzzz
1
0
Fork 0
Code Issues 6 Pull Requests Actions Packages Projects Releases Wiki Activity

Prepare for v2 development: Add MCP integration and future development planning

Browse Source 
- Add FUTURE_DEVELOPMENT.md with comprehensive v2 protocol specification
- Add MCP integration design and implementation foundation
- Add infrastructure and deployment configurations
- Update system architecture for v2 evolution

🤖 Generated with [Claude Code](https://claude.ai/code)

Co-Authored-By: Claude <noreply@anthropic.com>
This commit is contained in:
anthonyrawlins
2025-08-07 14:38:22 +10:00
parent 5f94288fbb
commit 065dddf8d5
41 changed files with 14970 additions and 161 deletions
Download Patch File Download Diff File Expand all files Collapse all files

16
UNIFIED_DEVELOPMENT_PLAN.md
View File

@@ -1,11 +1,11 @@
# Project Bzzz & Antennae: Integrated Development Plan
# Project Bzzz & HMMM: Integrated Development Plan
## 1. Unified Vision
This document outlines a unified development plan for **Project Bzzz** and its integrated meta-discussion layer, **Project Antennae**. The vision is to build a decentralized task execution network where autonomous agents can not only **act** but also **reason and collaborate** before acting.
This document outlines a unified development plan for **Project Bzzz** and its integrated meta-discussion layer, **Project HMMM**. The vision is to build a decentralized task execution network where autonomous agents can not only **act** but also **reason and collaborate** before acting.
- **Bzzz** provides the core P2P execution fabric (task claiming, execution, results).
- **Antennae** provides the collaborative "social brain" (task clarification, debate, knowledge sharing).
- **HMMM** provides the collaborative "social brain" (task clarification, debate, knowledge sharing).
By developing them together, we create a system that is both resilient and intelligent.
@@ -19,8 +19,8 @@ The combined architecture remains consistent with the principles of decentraliza
| :--- | :--- | :--- |
| **Networking** | **libp2p** | Peer discovery, identity, and secure P2P communication. |
| **Task Management** | **GitHub Issues** | The single source of truth for task definition and atomic allocation via assignment. |
| **Messaging** | **libp2p Pub/Sub** | Used for both `bzzz` (capabilities) and `antennae` (meta-discussion) topics. |
| **Logging** | **Hypercore Protocol** | A single, tamper-proof log stream per agent will store both execution logs (Bzzz) and discussion transcripts (Antennae). |
| **Messaging** | **libp2p Pub/Sub** | Used for both `bzzz` (capabilities) and `hmmm` (meta-discussion) topics. |
| **Logging** | **Hypercore Protocol** | A single, tamper-proof log stream per agent will store both execution logs (Bzzz) and discussion transcripts (HMMM). |
---
@@ -33,7 +33,7 @@ The agent's task lifecycle will be enhanced to include a reasoning step:
1. **Discover & Claim:** An agent discovers an unassigned GitHub issue and claims it by assigning itself.
2. **Open Meta-Channel:** The agent immediately joins a dedicated pub/sub topic: `bzzz/meta/issue/{id}`.
3. **Propose Plan:** The agent posts its proposed plan of action to the channel. *e.g., "I will address this by modifying `file.py` and adding a new function `x()`."*
4. **Listen & Discuss:** The agent waits for a brief "objection period" (e.g., 30 seconds). Other agents can chime in with suggestions, corrections, or questions. This is the core loop of the Antennae layer.
4. **Listen & Discuss:** The agent waits for a brief "objection period" (e.g., 30 seconds). Other agents can chime in with suggestions, corrections, or questions. This is the core loop of the HMMM layer.
5. **Execute:** If no major objections are raised, the agent proceeds with its plan.
6. **Report:** The agent creates a Pull Request. The PR description will include a link to the Hypercore log containing the full transcript of the pre-execution discussion.
@@ -74,7 +74,7 @@ This 8-week plan merges the development of both projects into a single, cohesive
| **1** | **P2P Foundation & Logging** | Establish the core agent identity and a unified **Hypercore log stream** for both action and discussion events. |
| **2** | **Capability Broadcasting** | Agents broadcast capabilities, including which reasoning models they have available (e.g., `claude-3-opus`). |
| **3** | **GitHub Task Claiming & Channel Creation** | Implement assignment-based task claiming. Upon claim, the agent **creates and subscribes to the meta-discussion channel**. |
| **4** | **Pre-Execution Discussion** | Implement the "propose plan" and "listen for objections" logic. This is the first functional version of the Antennae layer. |
| **4** | **Pre-Execution Discussion** | Implement the "propose plan" and "listen for objections" logic. This is the first functional version of the HMMM layer. |
| **5** | **Result Workflow with Logging** | Implement PR creation. The PR body **must link to the Hypercore discussion log**. |
| **6** | **Full Collaborative Help** | Implement the full `task_help_request` and `meta_msg` response flow, respecting all safeguards (hop limits, TTLs). |
| **7** | **Unified Monitoring** | The Mesh Visualizer dashboard will display agent status, execution logs, and **live meta-discussion transcripts**. |
@@ -84,4 +84,4 @@ This 8-week plan merges the development of both projects into a single, cohesive
## 5. Conclusion
By integrating Antennae from the outset, we are not just building a distributed task runner; we are building a **distributed reasoning system**. This approach will lead to a more robust, intelligent, and auditable Hive, where agents think and collaborate before they act.
By integrating HMMM from the outset, we are not just building a distributed task runner; we are building a **distributed reasoning system**. This approach will lead to a more robust, intelligent, and auditable Hive, where agents think and collaborate before they act.