# ElizaOS Weekly Newsletter (2026-02-22 → 2026-02-28) ## 1) Executive Summary **1. A real-world “regulated industry” plugin landed.** Community builder **Meme Broker** shipped a new **credit-building automation plugin** that goes beyond typical social/trading agents—supporting **FCRA/ECOA/FDCPA dispute letter generation**, **FICO analysis**, **business credit flows**, and even **Lob-certified mail sending**. It’s an exciting milestone for “agents that do paperwork,” but it also triggered an important compliance conversation. **2. Version stability got a practical path forward: `v2-develop`.** A confusing (and painful) repo state emerged where the **develop** branch unexpectedly contained **2.0.0** code instead of mature 1.x. The community response was decisive: create a **`v2-develop` branch** as the “safer 1.x landing zone” while **v2.0.0 remains alpha**. **3. Community momentum is picking up again.** In the **xfn-framework** channel, maintainers noted **external developer PRs are increasing**—a strong health signal—while also flagging the need to prevent **plugins “creeping back” into core** (see PR #6531). --- ## 2) Development Updates (Framework, Plugins, Fixes) ### Versioning & runtime reality (what to use today) - **v2.0.0 is still alpha**, and the community reiterated that **breaking changes are common** due to multiple runtime versions in active motion. - **Recommendation:** if you’re building on 1.x patterns and want fewer surprises, use **`v2-develop`** rather than the v2.0.0 release line right now. - A notable known issue: **bcrypt problems in v2.0.0** were mentioned as requiring patches. ### Plugin reliability concerns (pain points surfaced) - Developers reported that several common plugins are **broken out-of-the-box on 1.7.2**, including: - `plugin-linear` - `plugin-rolodex` - `plugin-memory` If you’re onboarding a team or planning a demo, budget time for patching—or align on a known-good branch/version as a group. ### Autonomy: three parallel implementations The community clarified there are **three “autonomy tracks”** in the ecosystem right now: 1. **`plugin-autonomous`** — enables *periodic thinking* so agents can run tasks on a schedule. 2. **v2.0.0 built-in autonomy** — “Shaw’s autonomous system” integrated into the newer runtime. 3. **The “milaidy” project** — described as more **OpenClaw-like**, aiming at deeper autonomy. Also noted: the **1.x tasks system** behaves like a cron, but **isn’t chat-configurable**, and folks are curious whether **`plugin-pim`** fills that gap. ### Repo hygiene & process - A version-control anomaly (develop branch containing 2.0.0 code) was described as effectively **untraceable via normal PR/commit history**, forcing the **`v2-develop`** workaround. - There was additional friction from **GitHub ↔ Linear bidirectional sync**, which created issue-tracking noise and cleanup work. --- ## 3) Community Spotlight (Discord Contributions & Discussions) ### Meme Broker’s credit-builder plugin: huge capability, bigger responsibility This week’s most talked-about contribution was the **credit-building plugin** (and its “plugin-form candidate” quality). It sparked two important threads: - **Product excitement:** agents handling certified mail and structured letter workflows is a major step toward “agents as operations staff.” - **Compliance questions (open):** **Caesar** asked how the plugin verifies **FCRA compliance** and prevents **improper disputes**—a critical unresolved topic. If you’re building in regulated spaces, this is the kind of “guardrails first” thinking we want to normalize. ### Enterprise workflow planning: how teams want to use ElizaOS **Julio Holon** outlined a realistic internal-ops agent roadmap: - Convert **Google Meet minutes** into **Linear issues** - Autonomously monitor **blocked issues** - Draft **PRs for human review** **Caesar** offered practical architecture guidance: - Use **persistent storage + embeddings** for memory - Add **human-in-the-loop confirmations** for high-stakes actions - Use **cron-like polling** (hourly) for reliable monitoring - Integrate via `plugin-linear`, `plugin-github`, `plugin-google` ### Security reminder (timely) **Omid Sa** warned that “ticket links” and DMs can be scams, clarifying they’re a community member—not staff. A good reminder: treat unexpected support outreach as hostile by default. ### Framework comparison: ElizaOS vs OpenClaw (production lens) A useful synthesis emerged: - **ElizaOS:** more stable, agent-framework oriented, and **safer for financial/enterprise data** - **OpenClaw:** broader “OS-like” autonomy (browser, sessions, nodes) but **frequent breaking changes** Choose based on your risk tolerance and required autonomy depth. --- ## 4) Token Economics (AI16z / $elizaOS + auto.fun) ### Token naming & migration clarity - A recurring confusion popped up: the correct token reference is **`$elizaOS` (not `$eliza`)**. - There was an **unanswered question** about whether **ai16z → elizaOS migration** remains possible **after the Feb 4 deadline**. If you have authoritative migration details (links, contracts, official instructions), now is a great time to share them in Discord so we can pin a definitive answer. ### auto.fun status No concrete **auto.fun** development updates were included in this week’s aggregated notes. If auto.fun has shipped changes, liquidity updates, or new mechanics this week, drop references/links in Discord so next week’s newsletter can cover them precisely. --- ## 5) Coming Soon (What to Watch Next) - **Compliance guardrails for regulated automation:** expect follow-up design discussions around **verification, consent, audit trails, and safe defaults**—especially for the credit-builder plugin category. - **Plugin maintenance & version compatibility push:** broken plugin reports (1.7.2) are now visible; likely next steps include compatibility matrices, pinned “known good” versions, or consolidated patch PRs. - **Autonomy bake-off:** community interest is rising in comparing `plugin-autonomous` vs v2 autonomy vs milaidy autonomy—especially around scheduling, chat configurability, and reliability. - **Babylon / milady.ai timing:** multiple users asked “when launch?”—no confirmed dates in the notes, but clearly top-of-mind. --- ## 6) Resources (Links & References) - **Credit Builder Plugin (GitHub):** https://opengraph.githubassets.com/e340e49d351d801e83724c5dbc7a868c5853e2d9268d3e9eba73c4920e1dbf0f/NewSoulOnTheBlock/plugin-credit-builder - **Discussion thread (credit-builder / coders):** https://discord.com/channels/1253563208833433701/1300025221834739744 - **xfn-framework note on plugin creep + PR reference:** - Discord: https://discord.com/channels/1253563208833433701/1377726087789940836 - PR (mentioned): https://github.com/elizaos/eliza/pull/6531 - **Cron Job Episode 4 + companion news site:** https://elizaos.news - **Open issue: URL messages causing duplicate LLM calls (cost + UX impact):** https://github.com/elizaos/eliza/issues/6486 - **Open feature request: custom OpenAI endpoint URL support:** https://github.com/elizaos/eliza/issues/6490 If you’re building this week: share your runtime version + plugin set when asking for help. The fastest fixes are the ones we can reproduce.