# ElizaOS Weekly Newsletter (2026-02-15 → 2026-02-21)

## 1) Executive Summary

This week felt like ElizaOS shifting into “beta-launch velocity,” with progress landing across core infra, security, and ecosystem growth:

1. **On-chain agent identity landed (SAID Protocol)**: the ecosystem added **SAID Protocol** support so agents can have a verifiable Solana identity, pushing ElizaOS closer to a composable, trustable agent economy.
2. **Workflow automation reliability improved (n8n + plugin ecosystem momentum)**: the **n8n workflow plugin** shipped a critical fix around default values/visibility checks, while the registry continued expanding with new integrations.
3. **Security tooling proved itself in the wild**: community-built detection tooling **flagged two Solana rugpulls (DARKTRUMP, NVIDIA) ~2 hours before crashes**, and discussions accelerated around cryptographic identity + trust scoring for agent-to-agent (A2A) interactions.

## 2) Development Updates (Core, Plugins, Infra)

### Core: identity + performance architecture
- **SAID Protocol integration for ElizaOS agents** is now in motion via a major PR:
  - **PR #6510**: *“feat: add SAID Protocol on-chain Solana identity for ElizaOS agents”* — adds Ed25519 keypair generation (Node `crypto`, no new deps) and registers an identity during `elizaos create`, saving a local wallet file and printing a public profile URL.
- **Database refactor work continues**:
  - **PR #6509**: *“chore: the great database refactor”* — focuses on separation of concerns (moving Drizzle out of core), unifying SQL interfaces, pagination support, upserts, and performance improvements (noted as meaningful speedups when handling multiple items). This is foundational work that should reduce friction for plugin authors and large deployments.

### Plugins: workflow automation + registry additions
- **plugin-n8n-workflow bug fix** (high impact for real agents using structured workflows):
  - Fix ensures **default property values are applied before visibility checks**, preventing dependent-field breakage when LLMs omit structural fields. In practice: fewer “phantom invalid configs” and less brittle node execution.
- **Registry expansion**:
  - New additions in progress/announced include **`@elizaos/plugin-said`** (SAID Protocol identity) and **SolCex Exchange BD plugin**, continuing the trend of finance + identity primitives becoming first-class citizens in the ecosystem.

### Developer experience: documentation + setup pain points
- **Spartan setup remains rough**: devs reported `bun install` hanging due to missing plugins (EVM, Farcaster, Jupiter, Knowledge, MySQL, Solana), requiring manual cloning. Docker files exist but are currently **not functional**. The silver lining: **plugin install order doesn’t matter**, which helps debugging.
- **Milady distribution gotcha**: `npx milady` installed an unrelated/unmaintained package (Alibaba tool). The project is **pre-release**; recommended path is **GitHub release binaries** for now. An issue was opened to address the npm naming confusion (**Issue #324** mentioned in Discord).

### Docs + collaboration tooling
- The team discussed adopting **HackMD MCP (Model Context Protocol)** tooling for shared notes and easier context handoff (reducing copy/paste across channels). Repo shared: `https://github.com/yuna0x0/hackmd-mcp`.

## 3) Community Spotlight (Discord highlights)

**jin — elizaos.news momentum**
- Continued improvements to **elizaos.news**, including better **mobile support** and design iteration (the “spread” layout got positive feedback).
- The platform’s pipeline is notable: **100% automated video generation** (PlayCanvas-based, not AI rendering), plus subtitles and agent narration.
- Coming concept teased: **1–1 interview shows** where a user chats with a bot, then the conversation becomes a generated interview episode (human + AI flavor).

**Vlt9 — real-world security validation**
- Shared live-test results showing a detection system that identified **DARKTRUMP** as *High-Risk* while trading near peak and also flagged the **NVIDIA token rugpull**—both about **2 hours before the crash**.
- Documentation was added, and **test keys** were offered to community members who want to validate/extend the system.

**Dawn (SageMind AI) — identity + trust layer for agents**
- Introduced **MoltBridge**, a cryptographic identity + trust/discovery layer for A2A interactions:
  - Uses **Ed25519 signatures** (no API keys), a **graph-based trust score**, and broker discovery.
  - SDKs available via npm/PyPI (`moltbridge`) and actively recruiting **~50 founding agents** to test integrations.
- Technical discussion explored combining **ERC-8004** on-chain anchoring with MoltBridge’s off-chain identity for speed + composability.

**ElizaBAO — partnerships, grants, ecosystem shipping**
- Reported meetings with **Solana Foundation** and **SuperteamMYR**, and noted **Clawlana submitted a Solana Grant application**.
- Confirmed **Shaw attended EthDenver** and participated in **ETHBuilder talks**.

## 4) Token Economics (AI16z, $elizaos, auto.fun)

- **$elizaos**: community Q&A reaffirmed **$elizaos is the main token**. The team stated **buybacks are funded with revenue**, with a commitment to clarify *timing and public disclosure* practices with Ops.
- **ai16z migration**: multiple users asked about missing the deadline; the answer was consistent and firm:
  - The **90-day migration window is over**, and community mods warned to **avoid scams** offering “late migration help.”
- **auto.fun**: no major new ship updates surfaced in the provided discussions this week. If you’re building around auto.fun, this is a good time to share what metrics and primitives you’d want surfaced in future updates (launch analytics, fee flows, incentive hooks, etc.).

## 5) Coming Soon (What to watch)

- **SAID identity becoming a default agent primitive**: expect follow-ups around opt-out controls, better key handling defaults, and how identity ties into registries/marketplaces.
- **Agent trust + security signal fusion**: MoltBridge + Security Oracle concepts are converging—watch for prototypes that map **Sybil detection / insider concentration** into a reusable trust graph.
- **Spartan onboarding improvements**: clearer “required vs optional plugins” docs and **working Docker** would materially lower contributor friction.
- **elizaos.news interview mode**: early storyboard + UX decisions are underway; this could become a flagship “community media” surface for agents.

## 6) Resources (Links & Pointers)

- Plugin creation guide (official): https://docs.elizaos.ai/guides/create-a-plugin  
- HackMD MCP integration tool: https://github.com/yuna0x0/hackmd-mcp  
- Core PRs:
  - SAID Protocol identity: https://github.com/elizaos/eliza/pull/6510  
  - Database refactor: https://github.com/elizaos/eliza/pull/6509  
- Known issues to track:
  - Ollama embeddings failing on Ubuntu 24.04: https://github.com/elizaos-plugins/plugin-ollama/issues/17  
  - Webapp duplicate LLM calls when message contains a URL: https://github.com/elizaos/eliza/issues/6486  
  - Custom OpenAI endpoint request: https://github.com/elizaos/eliza/issues/6490  

If you want to help next week: jump into **workflow automation (n8n)** testing, volunteer as a **founding agent** for identity/trust experiments, or pick off a **Spartan setup/doc** improvement that reduces time-to-first-agent for new builders.