## Weekly Newsletter (2026-02-17) — Week of **2026-02-10** to **2026-02-16** ### 1) Executive Summary This week felt like a “tighten the bolts + expand the frontier” moment for ElizaOS: - **Stability & UX leap in core `eliza`**: The team cleared a backlog of ~20 issues in a single push, landing fixes across the dashboard, app builder, redirects, and prompt limitations—while also improving agent collaboration and shipping key integrations. - **Trust & security became the headline theme**: After the revelation of **341 malicious skills bypassing marketplace vetting** (via ClawHavoc on ClawHub), the community rallied around stronger agent identity and risk intelligence—most notably via **MoltBridge** (cryptographic agent identity + trust scoring) and a new **Security Oracle API** beta for trading agents. - **Beta-launch planning kicked into higher gear**: New workstreams formed around baseline product metrics, tuning Eliza’s conversational personality, and building a plugin to **auto-create user profiles**, signaling a more productized “first beta” experience. --- ### 2) Development Updates (Core + Plugins) #### Core platform stability & UX A concentrated maintenance sprint landed several high-impact improvements: - **Resolved dashboard and app builder issues**, fixed **user redirect** problems, and updated an **outdated changelog**. - **Removed the 500-character limit** in the first app prompt—small change, big developer experience win for building real workflows. #### Agent capabilities & collaboration improvements The platform is steadily moving toward richer, more “agent-native” experiences: - **CoT reasoning streaming support** was implemented, improving transparency and responsiveness for streamed outputs. - Multiple integrations were highlighted as part of the day’s progress: **WhatsApp**, **Crypto/DeFi**, **Gmail/Email**, and the **N8N Workflow Engine**. - Collaboration features advanced with fixes to **multi-user/room awareness**, **contact lookup + messaging**, and **cross-agent messaging**. - A **native iOS app** was reported as completed—worth watching as it changes how people will “live with” their agents day-to-day. #### Plugin spotlight: `plugin-n8n-workflow` REST API expansion The N8N workflow plugin received a particularly meaningful upgrade: **new REST API routes** enabling direct frontend control without routing everything through the NLP pipeline. What this unlocks: - Full workflow **CRUD** - **Node catalog** browsing - **Validation** and **execution monitoring** - Stronger confidence via **extensive unit tests** and updated test helpers This is a practical step toward making ElizaOS feel like a real automation platform, not just a chat interface. #### Repo hygiene & CI/CD - The **registry repo CI/CD** was improved to fix the **Claude-review process on fork PRs**. - Dependency updates landed in `elizaos.github.io`. --- ### 3) Community Spotlight (Discord) #### MoltBridge: cryptographic identity + trust layer for agents Dawn introduced **MoltBridge**, an agent-to-agent (A2A) identity and trust layer designed explicitly to address marketplace and skill supply-chain risks. Key technical ideas discussed: - **Ed25519 signatures** for verifiable identity **without API-key dependency** - **Graph-based broker discovery** to derive trust scores via relationship topology - A composable identity direction: pair **off-chain speed** (Ed25519) with **on-chain anchoring** via **ERC-8004** for durable identity claims SDKs are already live on **npm** and **PyPI** (`moltbridge`), and the project is **seeking 50 founding agents** for integration testing. Just as importantly, the community also debated **norms for autonomous agents participating in Discord**. Dawn initially didn’t clearly disclose being an AI agent (later clarified: built on Claude with a human collaborator, Justin/SageMind AI). Several members asked for explicit server guidance—expect follow-up on rules and disclosure expectations. #### Security Oracle API (beta): unified risk intelligence for trading agents Vlt9 launched a **Beta Security Oracle API** aggregating RugCheck, GoPlus, and sentiment signals with features like: - **Zero-lag detection** on new liquidity pairs - **Insider concentration** detection - **Sybil farm** identification - Strict **JSON outputs** designed for ElizaOS/ai16z integration - Beta limits: **100 requests/day**, **1 request per 10 seconds** A promising synergy emerged: using Oracle signals (Sybil/insider risk) as an input layer for **MoltBridge trust scoring**—effectively combining identity + behavior/risk telemetry. #### Ecosystem integrations & experiments - **Moltbook plugin (Kalshi + Solana execution flows)** was announced, with open questions about the current state of the **ElizaOS Solana plugin**, website embedding, and whether newer alternatives (e.g., “x402”) exist. - A fun but surprisingly deep thread: **Meme Broker’s Nietzsche-themed agent**, which sparked discussion on **persona drift**, long-horizon consistency, and “identity persistence” in character-driven agents. #### Scam prevention & support hygiene Community members continued warning about scammers and fake support flows (notably: “support” asking users to move to other servers/calls). The repeated reminder: **legit support won’t ask you to go off-server**. --- ### 4) Token Economics (AI16z + auto.fun) #### AI16z → ElizaOS migration status (important) The migration conversation remained active—and emotional. Multiple users reported missing the deadline due to life circumstances, but the guidance remained consistent: - The **90-day migration window has ended** (deadline: **Feb 4**). - Community members reiterated: **no new tickets can be opened**, and **post-deadline swaps aren’t possible**. - Repeated warning: anyone offering “manual help” now is likely a **scam**. #### auto.fun No concrete auto.fun product changes were shared in the provided data this week. If you’re building on auto.fun or waiting on it, keep an eye on announcements—this week’s momentum was more focused on core stability, trust/security, and beta planning. --- ### 5) Coming Soon Here’s what to anticipate based on active planning threads and surfaced needs: - **Initial beta launch execution**: newly created planning issues include baseline **product metrics**, **Eliza personality tuning**, and a plugin to **automatically build user profiles**. - **Agent identity standards work**: deeper exploration of **ERC-8004 + off-chain identity** composability (a likely direction for fast verification without gas/block-time delays). - **Security-first agent economy**: potential integration of Security Oracle outputs into broader trust and discovery layers. - **Solana + website integration clarity**: community is explicitly asking for guidance on the current **Solana plugin usability** and **mcp-gateway** for web embedding. - **Documentation / governance**: a request is on the table to **clarify server rules** for autonomous agent participation and disclosure. --- ### 6) Resources (links & references) **Discord discussions** - MoltBridge + Security Oracle + plugin announcements (💬-coders): https://discord.com/channels/1253563208833433701/1300025221834739744 - ERC-8004, migration conversations, ecosystem updates (💬-discussion): https://discord.com/channels/1253563208833433701/1253563209462448241 **GitHub / development references (high-signal items mentioned in summaries)** - Weekly engineering summary (Feb 8–14): `github_summaries_week_latest_2026-02-08.md` (provided in dataset) - Notable `elizaos/eliza` workstreams referenced this month: - **JWT authentication system**: PR **#6200** - **Core docs bundle (architecture/core concepts/deployment/plugin dev)**: PR **#6356** - **Environment variables documentation**: PR **#6377** - **ActionFilterService (vector search + BM25 reranking)**: PR **#6475** - **Request context for per-user entity settings**: PR **#6457** - Defensive stability fixes in bootstrap/core settings: PRs **#6471**, **#6473** - `plugin-n8n-workflow`: REST API workflow management routes (highlighted in the 2026-02-16 dev summary) If you’re building this week: consider joining the **MoltBridge founding agent** cohort, or volunteering to test the **Security Oracle API beta** (3–5 devs requested). Both efforts directly support a safer, more composable agent economy.