# ElizaOS Weekly Newsletter (2026-01-20 to 2026-01-26)

## 1) Executive Summary

**1. Eliza Cloud unblocked + new core capabilities landed.**  
A blocking issue with **AI Gateway credits** was resolved, getting inference and day-to-day development moving again. Alongside that, the team integrated foundational plugins—most notably a **GitHub plugin** (repo/code management) and a **Code Execution plugin** (agents can write + run code), pushing Eliza closer to a “hands-on” agent platform rather than just chat.

**2. V1.7.2 stabilization momentum (CLI + plugin ecosystem).**  
A critical CLI fix updated an import to use the proper package alias, improving module resolution and preventing “can’t create project” style breakages. On the community side, 1.7.2 upgrade friction (migrations + cached versions) surfaced and was actively troubleshot, with clear remediation steps shared.

**3. Ecosystem strategy took center stage (tokenomics + distribution).**  
Discord saw an intense, high-signal debate about **ecosystem token launches** (notably Hyperscape’s “gold” token), funding runway realities, and how to align new app tokens with **$ELIZAOS** value. The thread produced actionable proposals (fees/burns, airdrops, pairing requirements, grants) and clarified what the team is optimizing for.

---

## 2) Development Updates

### Core fixes & infrastructure
- **CLI import bug fix:** The CLI entrypoint was updated to import via the package alias (instead of a deep path), improving compatibility with modern Node/Bun module export rules. This addresses a class of failures where `elizaos create` would crash due to “subpath not exported” errors.
- **AI Gateway credits restored:** A previously blocking credits issue was resolved, unfreezing inference and enabling developers to test agent behavior reliably again.
- **Model naming convention flagged as urgent:** An internal/documentation mismatch emerged around model tier naming—specifically the mapping between **Opus/Sonnet** and **Ultra/Pro** labels. This impacts docs, configuration, and shared understanding, so expect clarifications (and possibly config defaults) soon.

### Plugin & agent capability expansions
- **GitHub Plugin integrated:** Enables agents to reason about repositories and manage code workflows more directly.
- **Code Execution Plugin integrated:** Expands agent autonomy by letting Eliza write and execute code as part of tasks—useful for debugging, data transforms, and prototyping.

### Web + docs maintenance
- **Dependency updates across `elizaos.github.io`:** A broad set of minor/patch bumps (e.g., `next`, `commander`, `date-fns`, `glob`, `ora`, `eslint-config-next`) landed to keep the site secure and current.

### Roadmap signals (new issues opened)
Engineering also sketched a big forward slate, including:
- **Poke-style onboarding**, **conversation-based configuration**, and a **landing page**
- Integrations for **Discord, Telegram, SMS/iMessage, Twilio, WhatsApp**
- **Dynamic plugin loading**, **multi-tenant session isolation**, **secrets management**
- **Memory persistence**, **cross-agent messaging**
- Longer-horizon: **native iOS app development**

---

## 3) Community Spotlight (Discord)

### MCP integration: connecting external tools to Eliza Cloud
A standout technical thread explored using **Model Context Protocol (MCP)** servers with Eliza Cloud—particularly for **AI red teaming** with human-in-the-loop workflows. The key takeaway: you *can* register external MCP servers via the Cloud API and let Eliza proxy calls during reasoning and task execution.

- Practical detail shared: `POST /api/v1/mcps` registers an external MCP endpoint for agent use.
- Community interest: adding “Eliza MCP functionality” more broadly (and making it easier to discover/use) was explicitly requested.

### Debugging: “plugin action handler callbacks” behavior
In `💬-coders`, Victor Creed reported a confusing behavior in **v1.7.2**: only the first callback message appeared, while docs imply multiple callbacks are supported and should feel “immediate” (not only at action completion).  
Odilitime confirmed that **multiple callbacks are supported** and began troubleshooting—starting with task planner configuration (onestep vs multistep). This is a good example of community-led validation of framework guarantees, and it likely points to either configuration nuance or a doc clarity gap.

### Real-world deployment patterns: mini PCs + free-tier cloud
A practical self-hosting thread highlighted a pattern: community members are running Eliza on **mini PCs** to avoid keeping main machines on 24/7. An **Oracle Cloud free tier** configuration was referenced (notably **4 vCPU / 24GB RAM**) as a viable starting point for budget hosting.

### Security awareness: scams are actively targeting users
Two scam patterns came up:
- A “Create A Ticket” bot requesting wallet addresses (flagged as malicious).
- A fraudulent “token migration support” DM instructing users to send tokens to a wallet for a promised swap.  
Takeaway: **no legitimate migration requires sending tokens to an unsolicited wallet address**. Use official channels and the ticket system only.

### People updates
- **Sam returned** from medical leave (80% recovered) and is resuming cloud infrastructure work.
- **Shaw rejoined** the server (and participated in the larger ecosystem strategy discussions).

---

## 4) Token Economics (AI16z / $ELIZAOS / auto.fun)

### Price sentiment & communication needs
Community sentiment was tense around continued **token price weakness** (new ATLs). Several members asked bluntly whether $ELIZAOS remains “valid” and wanted clearer team communication during the downturn. Others reinforced a long-term thesis: AI agents are early, markets are broadly down, and ElizaOS remains a leading project in the category.

### Migration support & risk
Migration questions (Ledger, **ai16z → elizaOS**) were frequent—and unfortunately coincided with scam attempts. The consistent guidance from moderators/community was to route users through the **official ticket channel** for support, not DMs.

### Ecosystem launches & auto.fun dynamics
The Hyperscape “gold” token debate effectively became a proxy for how the community wants **auto.fun-style launches** (and similar experiments) to connect back to $ELIZAOS. Strong proposals emerged:
- **Airdrops to $ELIZAOS holders** for ecosystem tokens
- **Pairing requirements** (ecosystem tokens paired with $ELIZAOS, Virtuals-style)
- **Grants-based funding** rather than team-launched tokens
- **Direct utility linkage:** use $ELIZAOS for platform fees (LLM compute/storage) and/or burns triggered by in-game item/agent creation

---

## 5) Coming Soon

Here’s what to watch next week(s), based on active discussion and newly opened work:
- **Clarification/fix for model naming mappings** (Opus/Sonnet vs Ultra/Pro)
- **Follow-through on callback behavior investigation** (and likely doc updates)
- **More MCP ergonomics** (potentially broader visibility + easier setup)
- **Cloud apps launch “imminent”** (per team commentary), plus Hyperscape MVP progress
- **New onboarding + configuration flows** to reduce friction for first-time builders
- Potential partnership exploration: **Seeker phone DApp store integration** (Eliza as a Claude-style assistant app)

---

## 6) Resources

- **External MCP registration (Eliza Cloud API):** `POST /api/v1/mcps` (shared in Discord)  
- **DaVinci Resolve MCP integration repo:** https://github.com/AyeRubz/davinci-resolve-mcp  
- **Token migration / support tickets:** Discord channel **#1425417640071139358**  
- **Weekly strategic debate (tokenomics + ecosystem funding):** Discord logs from **2026-01-24** (discussion highlights)  
- **CLI / v1.7.2 upgrade troubleshooting steps:** shared by **0xbbjoker** (cache + reinstall + clean lockfiles/node_modules)  
- **Seeker integration discussion:** Discord `🥇-partners` thread (2026-01-26)