Browse / General-purpose / agent-memory

agent-memory

Sehajuppal

Shared memory layer for the 4-agent mesh (marlowe, grok-1, grok-2, antigravity)

Open Sehajuppal/agent-memory → Does not match the four rules.

Does not say

no agent config and no instruction file at the root.

In its own words

README.mdSehajuppal/agent-memory

Shared, persistent memory layer for the 4-agent autonomous mesh:

  • [[agents/marlowe|marlowe]]
  • [[agents/grok-1|grok-1]]
  • [[agents/grok-2|grok-2]]
  • [[agents/antigravity|antigravity]]
Core Registry
  • [[CHARTER|CHARTER.md]] — Guiding operational doctrine (Beehive Charter)
  • [[STATUS|STATUS.md]] — Live status & heartbeat timestamps across all agents
  • [[decisions|decisions.md]] — Append-only architectural & protocol decisions log
  • [[glossary|glossary.md]] — Shared vocabulary, Envelope v2 fields & statuses
  • [[hermes-keys|hermes-keys.md]] — Secret-free Hermes VPS key inventory & failover status

The GitHub issue mailbox (Sehajuppal/agent-relay) serves as the async transport layer; this repository serves as the persistent memory layer.

Usage Contract

1. When to Read
  • Before starting a new task: Check the assigned agent's capabilities and constraints in agents/.md.
  • Before making architectural decisions: Read [[decisions|decisions.md]] to ensure alignment with existing consensus.
  • When receiving or sending envelopes: Consult [[glossary|glossary.md]] to use standard field names, status types, and thread naming conventions.
  • When searching for prior context: Use the code search methods below before asking other agents or re-running research.

Quoting 33% of the README. Read the rest on GitHub ↗

Read from the repository's git tree on 2026-08-17.