## Intel Brief — 2026-04-10 (ElizaOS)

### 1) Data Pattern Recognition (Quant + Trends)

**Development velocity (rolling month 2026-04-01 → 2026-05-01, elizaos/eliza)**
- PRs: **4 opened / 3 merged**
- Issues: **7 opened / 1 closed**
- Contributors: **14 active**
- Code churn (latest daily snapshot): **70 commits**, **+6,785 / −3,732** lines across **81 files**
- Concentration risk: top contributor **odilitime** leads by a wide margin (highest total score; also authored/merged multiple core PRs). Bus factor remains a concern for core runtime changes.

**Core/platform trajectory**
- Clear push toward **runtime hardening + dev ergonomics**:
  - New **agent workspace** (stdin/stdout harness) to streamline local boot.
  - **Multi-language support** landing in core (Python/Rust).
  - Runtime-composition APIs refactored to support flexible **character loading**.
- **Plugin ecosystem pressure increasing**:
  - Anthropic plugin updated for **Opus 4.x** compatibility (ai-sdk v6 param changes).
  - Discord plugin chasing feature parity (typing, reactions, streaming, slash commands).
  - Multiple new plugin proposals/issues (marketplace, safety, incentives, “economic memory”).

**Community engagement patterns (Discord, last 3 days)**
- Engagement skewed toward:
  1) **Deployment simplification** (Hatcher.host repeatedly promoted + validated by a successful deployment report).
  2) **Project/hype discovery** (elizabao_ai “what is it / what can I do now?” repeated).
  3) **Market sentiment** (token price drawdown discussion; negative tone and distraction risk).
- Moderation/security signal: repeated **scammer flags** in chat (4/07), suggests heightened need for trust & safety comms.

**Pain-point correlation across channels**
- **Onboarding friction** correlates with both:
  - Tooling issues (e.g., CLI create failure on macOS due to bun dependency behavior).
  - Knowledge gaps (elizabao_ai unclear, agent safety question unanswered).
- **Safety concerns** are showing up in both community questions and technical direction (AgentID/capability authorization discussions), indicating a converging priority.

---

### 2) User Experience Intelligence (Impact + Themes)

**Theme A — “What is elizabao_ai and what can I do today?” (High impact, high urgency)**
- Signal: direct unanswered user question in Discord (cryptovalutchik_); creator notes rising inbound interest.
- UX risk: hype without a clear “current value” narrative increases churn and mistrust.
- Opportunity: a tight “What it is / What’s live / What’s next / How to try” page plus a pinned Discord message.

**Theme B — No-code deployment demand (Medium-high impact, medium urgency)**
- Signal: Hatcher.host positioned as “deploy without code/servers/hardware”; a user reports successful 1-click deploy, clean logs/dashboard, Groq integration.
- Adoption implication: deployment tooling is becoming a primary funnel for new users; ElizaOS should treat it as an “official-ish” path even if third-party.
- Opportunity: publish an “approved deployment paths” doc section (Hatcher + self-host), with capability/safety caveats.

**Theme C — Agent safety / preventing unsafe operations (High impact, unresolved)**
- Signal: explicit question raised (4/08) with no resolution; multiple parallel GitHub proposals around capability enforcement and scoped authority.
- UX risk: without a simple “safety model” explanation, users will either over-trust agents or avoid production usage.
- Opportunity: unify messaging around GuardrailsService + a recommended pattern for “high-risk tool calls require approval”.

**Theme D — Community sentiment drag from token price talk + scams (Medium impact, ongoing)**
- Signal: extended token price negativity (4/07) and scammer callouts.
- UX risk: support and dev channels get diluted; newcomers may conclude project is purely speculative.
- Opportunity: tighten channel routing + enforce “support-first” surfaces; publish scam-awareness guidance + official link hygiene.

---

### 3) Strategic Prioritization (Impact × Risk × Dependencies)

#### Priority 0 (Blockers / Reputation Risk)
1) **elizabao_ai clarity pack (Docs + product surface)**
   - User impact: **Very high** (top repeated question; hype-driven).
   - Tech risk: **Low**.
   - Deliverables (48–72 hours):
     - 1-page doc: *What is elizabao_ai?* + “What can I do now?” + “How to get access / try”.
     - Pinned Discord FAQ + link to canonical doc.
     - If not usable yet: explicitly label as “in development” and provide a waitlist / preview flow.

2) **Agent safety baseline guidance (minimum viable policy)**
   - User impact: **High** (direct concern; aligns with AgentID direction).
   - Tech risk: **Medium** (touches runtime/tooling practices).
   - Deliverables (1 week):
     - Document a default **risk-tier model** for tools (observe/act/commit).
     - Recommend a standard “human approval hook” for commit-tier actions.
     - Provide a reference implementation (even if minimal) for gating tool calls.

#### Priority 1 (Developer Experience / Adoption)
3) **Fix/mitigate macOS `elizaos create` failure (Issue #6704)**
   - User impact: **High** (hard stop at first-run; CLI cleans up project dir, compounding frustration).
   - Tech risk: **Low–Medium** (dependency graph changes).
   - Action:
     - Remove `bun` as runtime dependency; keep only types as dev dependency (per issue writeup).
     - Update CLI to avoid deleting the project directory on build failure; preserve logs/artifacts.

4) **Stabilize core group-room routing before broad release (PR #6712)**
   - User impact: **Medium-high** (multi-party behavior determines perceived “intelligence” and reduces spam loops).
   - Tech risk: **Medium** (shouldRespond path changes).
   - Dependency: fix identified P1 bug in `aliasEntity`/token ambiguity; add tests that would have caught it.

#### Priority 2 (Ecosystem growth, but gate by quality)
5) **Plugin intake hardening (registry + review rubric)**
   - User impact: **Medium** (more capabilities), but quality failures harm trust.
   - Tech risk: **Medium**.
   - Immediate need: proposals like **plugin-mnemopay** show major gaps (no persistence, NaN corruption, null deref risks, no tests).
   - Action:
     - Introduce a “minimum bar” checklist: persistence expectations, config validation, bounded memory growth, required tests, failure modes.

---

## Actionable Recommendations (Next 7 Days)

**A) Ship “Discovery + Trust” artifacts**
- Publish:
  - `elizabao_ai.md` (what it is, current state, how to try).
  - “Safe agent operations” guide (tool risk tiers + approval patterns).
- Pin both in Discord; add a short “official links” message to reduce scam surface.

**B) Protect onboarding**
- Patch CLI create flow (macOS bun dependency + failure cleanup behavior).
- Add an onboarding “known issues” page until patch is released.

**C) Focus engineering on closing critical-path PR risk**
- Allocate review bandwidth to:
  - PR #6712 bugfix + tests (routing correctness).
  - PR #6709 (toon params + async terminal actions) if it’s gating connector reliability (Discord/Milady).

**D) Community operations hygiene**
- Create a lightweight “promo/tickets” intake mechanism (even a GitHub Discussions category or a Discord form) to prevent ad-hoc coordination overhead.
- Strengthen moderation playbook for scam detection + rapid takedown guidance.

---

## KPI Watchlist (to track starting next report)
- **Time-to-first-successful-deploy** (self-host vs Hatcher).
- **CLI create success rate** (macOS/Windows/Linux).
- **% unanswered “what is X?” questions** in Discord (elizabao_ai currently a hotspot).
- **Incidence of loop/spam behavior** in group channels (post addressee routing changes).
- **Plugin PR acceptance rate vs. rejection reasons** (tests/persistence/config validation).