chore: fill USERNAME placeholder with MakiDevelop before public release

Author: maki

README.md

@@ -13,7 +13,7 @@ Existing AI memory tools (mem0, Letta, …) are Claude-only or chat-only and bur
 - **One memory hall, many agents.** Claude writes via MCP, Codex writes via CLI, your shell scripts write via curl. They all read the same store.
 - **Cross-session by default.** Every entry is timestamped, agent-tagged, and namespaced. Pick up tomorrow where any agent left off today.
 - **Zero-LLM write path.** Writes are SQLite + embedding only — P99 < 100ms on local-only deployments (Ollama same host, sqlite-vec, single tenant). Optional enrichment runs in the background, never blocks you.
-- **Min local, max distributed.** Runs on a single laptop with `sqlite-vec` and local Ollama embeddings. Scales to a multi-host home AI lab with Qdrant + remote embedder via [memory-gateway](https://github.com/USERNAME/memory-gateway).
+- **Min local, max distributed.** Runs on a single laptop with `sqlite-vec` and local Ollama embeddings. Scales to a multi-host home AI lab with Qdrant + remote embedder via [memory-gateway](https://github.com/MakiDevelop/memory-gateway).
 - **Multi-tenant data model from day one.** Schema-ready for multi-user release; v0.1 runtime ships single-tenant for focus (see [ADR 0002](docs/adr/0002-multi-tenant-from-day-one.md)).
 
 ## Quickstart (single-agent, 1 minute, no auth)
@@ -22,7 +22,7 @@ Requires: `python>=3.12`, `ollama` running locally, `bge-m3` model pulled.
 
 ```bash
 ollama pull bge-m3
-git clone https://github.com/USERNAME/memory-hall
+git clone https://github.com/MakiDevelop/memory-hall
 cd memory-hall
 uv sync
 uv run memory-hall serve  # listens on :9000, dev mode (no auth, localhost-only)
@@ -33,7 +33,7 @@ curl -X POST http://localhost:9000/v1/memory/write \
   -d '{"agent_id":"shell","namespace":"home","type":"note","content":"hello memory hall"}'
 ```
 
-> **Auth note:** the standalone `memory-hall serve` runs in dev mode by default — no HMAC, no nonce, listens on localhost. For production / multi-user / network-exposed deployments, deploy via [`memory-gateway`](https://github.com/USERNAME/memory-gateway), which adds HMAC auth, ACL, and governance APIs on top of memory-hall. See [ADR 0003](docs/adr/0003-engine-library-vs-deployment-platform.md) for the engine/platform split.
+> **Auth note:** the standalone `memory-hall serve` runs in dev mode by default — no HMAC, no nonce, listens on localhost. For production / multi-user / network-exposed deployments, deploy via [`memory-gateway`](https://github.com/MakiDevelop/memory-gateway), which adds HMAC auth, ACL, and governance APIs on top of memory-hall. See [ADR 0003](docs/adr/0003-engine-library-vs-deployment-platform.md) for the engine/platform split.
 
 ## Quickstart (multi-host, your home AI lab)
 

docs/adr/0003-engine-library-vs-deployment-platform.md

@@ -5,7 +5,7 @@
 
 ## Context
 
-The author already maintains [memory-gateway](https://github.com/USERNAME/memory-gateway): a FastAPI service with HMAC auth, ACL, staging, proposals, and event-sourced governance APIs. Started in early 2026, paused mid-flight when its mem0-backed proxy turned out to inherit all of mem0's reliability problems.
+The author already maintains [memory-gateway](https://github.com/MakiDevelop/memory-gateway): a FastAPI service with HMAC auth, ACL, staging, proposals, and event-sourced governance APIs. Started in early 2026, paused mid-flight when its mem0-backed proxy turned out to inherit all of mem0's reliability problems.
 
 Now building memory-hall as the replacement memory layer raises an obvious question: **build into memory-gateway, or fork a new repo?**
 

docs/design.md

@@ -41,7 +41,7 @@ memory-hall is the **engine library**:
 - Storage, embedder, search, CLI, optional minimal FastAPI server for solo use
 - Pip-installable, runs without auth in local-only setups
 
-[`memory-gateway`](https://github.com/USERNAME/memory-gateway) is the **deployment platform** that wraps memory-hall with HMAC auth, ACL, staging, proposals, and event-sourced governance.
+[`memory-gateway`](https://github.com/MakiDevelop/memory-gateway) is the **deployment platform** that wraps memory-hall with HMAC auth, ACL, staging, proposals, and event-sourced governance.
 
 Solo users: `memory-hall serve` is enough. Production / multi-user: deploy `memory-gateway`, which embeds memory-hall as a library.
 

pyproject.toml

@@ -55,10 +55,10 @@ memory-hall = "memory_hall.cli.main:app"
 mh = "memory_hall.cli.main:app"
 
 [project.urls]
-Homepage = "https://github.com/USERNAME/memory-hall"
-Documentation = "https://github.com/USERNAME/memory-hall/tree/main/docs"
-Repository = "https://github.com/USERNAME/memory-hall"
-Issues = "https://github.com/USERNAME/memory-hall/issues"
+Homepage = "https://github.com/MakiDevelop/memory-hall"
+Documentation = "https://github.com/MakiDevelop/memory-hall/tree/main/docs"
+Repository = "https://github.com/MakiDevelop/memory-hall"
+Issues = "https://github.com/MakiDevelop/memory-hall/issues"
 
 [build-system]
 requires = ["hatchling"]