## ElizaOS Strategic Intel — 2025-12-25

### Executive Signal (24–25 Dec window)
- **Engineering throughput (GitHub /eliza):** 0 PRs merged, **4 new issues**, **2 active contributors** → **low merge velocity**, likely holiday-driven, but backlog is shifting toward **UX correctness + docs quality**.
- **Community focus (Discord):** Discussion heavily concentrated on **token value accrual clarity** and **Cloud capability boundaries** (uploads, custom models), with recurring requests for **canonical documentation**.
- **Operational fragility:** A CI job failed due to **Claude billing top-up** need → non-technical dependency is blocking automation.

---

## 1) Data Pattern Recognition

### Development velocity & trend
- **Short-term (Dec 24–25):** Merge velocity stalled (0 merges), but **issue intake increased** (4 opened) indicating **QA/feedback continues even during low dev activity**.
- **Month context (Dec overall, provided):** 28 PRs opened / 16 merged, 60 new issues, 22 contributors → **healthy baseline**, so the current dip is likely temporal rather than structural.

### Community engagement patterns
- Engagement clusters around:
  1) **Tokenomics/value accrual** (fear/uncertainty due to reported ~99.5% drawdown from peak; repeated “why invest?” loop)
  2) **Cloud roadmap questions** (custom model hosting timing; upload constraints; recommended formats)
  3) **Migration security** (scam warnings; wallet guidance)
  4) **Dev-experience friction** (CLI install/version mismatches; outdated monorepo docs)

### Feature adoption / demand signals (proxy from questions)
- **Knowledge upload demand** is high: users want to ingest large corpora (e.g., “300 pages of book data”).
  - Current constraint: **50MB upload limit**, **Markdown preferred**.
- **Custom model hosting** is a near-term expectation (“next week”) → high likelihood of disappointment if it slips without proactive comms.

### Pain-point correlation across channels
- **Docs debt ↔ onboarding friction:** Outdated monorepo docs + Bun/Node/NPM inconsistency reports + CLI install help requests form a single funnel problem: **first 30 minutes fail rate**.
- **Agent chat UX issues ↔ trust:** Conversation duplication between agents + not opening latest conversation by default → **perceived data isolation/safety risk** even if it’s “just UI/state.”

---

## 2) User Experience Intelligence

### Feedback themes (categorized by impact)

**High impact (blocks use / trust)**
- **Conversation duplication across agents** (Issue #6282): users seeing one agent’s chat appear under another → perceived privacy/data isolation breach.
- **Migration security anxiety:** repeated scam warnings; advice to move assets to Phantom for safer migration. This is a trust & safety surface, not “just support.”

**Medium impact (degrades adoption / conversion)**
- **Agent chat entry behavior** (Issue #6281): opening an agent from Dashboard should default to **most recent conversation**.
- **Cloud upload constraints clarity:** 50MB limit is clear once stated, but not discoverable enough; users are testing boundaries with large documents.
- **Custom model hosting expectation management:** “next week” promise creates a near-term expectation spike.

**Low impact (polish / paper cuts)**
- **Markdown/numbered list spacing issues** (Issue #6283) and general formatting consistency.

### Usage patterns vs intended design (inferred)
- Users are treating Cloud agents as **knowledge-base assistants** (book ingestion) rather than small “prompt-only” agents → product needs **knowledge management UX** (chunking, indexing status, retrieval transparency).
- Community expects token value accrual to be **explicit and documented**, not inferred from “usage drives token.”

### Sentiment tracking (qualitative)
- Sentiment is **bifurcated**:
  - Positive: holiday camaraderie; builders shipping (Bun frontend, plugin updates).
  - Negative: token drawdown frustration + uncertainty around token utility beyond buybacks.
- Risk: if tokenomics clarity doesn’t improve before January ramp-up, **Cloud launch narrative may be overshadowed**.

---

## 3) Strategic Prioritization

### Priority stack (Impact × Risk × Dependency)

#### P0 — Trust & Data Integrity (ship immediately)
1) **Fix conversation duplication across agents (Issue #6282)**
   - **User impact:** Very high (trust-breaking, “data leakage” perception)
   - **Tech risk:** Medium (state/session routing; caching; server/client sync)
   - **Dependency:** Ensure message/session identity and agent scoping are consistent across UI + API.

2) **Migration safety hardening (comms + UX guardrails)**
   - Add **official migration safety checklist** and “never send tokens to…” warnings in canonical docs and pinned Discord messages.
   - Consider a **signed/verified links page** (single source of truth).

#### P1 — Activation & Retention (next sprint)
3) **Docs correctness + onboarding consistency**
   - Close the loop on outdated monorepo docs (Issue #6284) + Bun/Node/NPM version guidance.
   - Target metric: reduce “install/troubleshoot” support interactions by **30–50%**.

4) **Agent chat UX: open most recent conversation (Issue #6281)**
   - Improves perceived polish and reduces confusion/support load.

#### P2 — Platform capability expansion (manage expectations)
5) **Custom model hosting (roadmap: “next week”)**
   - If delivery risk exists, publish scope boundaries now (supported runtimes, auth, quotas, latency expectations).
   - Avoid “silent slip” during January PR ramp.

### Critical path dependencies to watch
- **CI reliability dependency:** Claude billing top-up is blocking GitHub Actions; this is a **single-point operational failure** affecting velocity.
- **Docs + product claims alignment:** Token buyback mechanics and “yield vs revenue” distinctions must be consistent across Discord answers, docs, and any Cloud marketing pages.
- **Data isolation perception:** Any chat/session bug will undermine future multi-tenant/auth work and Cloud positioning.

---

## Quantitative Summary (today’s actionable numbers)
- **GitHub (/eliza) 24–25 Dec:** **0 merged PRs**, **4 new issues**, **2 contributors active**
- **Top user-requested Cloud constraints:** **50MB upload limit**, **Markdown preferred**, **custom model hosting expected ~1 week**
- **Dominant discussion domain:** token value accrual clarity (multiple threads across channels)

---

## Recommendations (actionable, execution-oriented)

1) **Ship a “Trust Patch” mini-release**
   - Fix **Issue #6282** (conversation duplication) with regression tests covering:
     - agentId/roomId/sessionId scoping
     - switching agents rapidly
     - concurrent tabs
   - Add a short release note explicitly stating “no cross-agent chat leakage” to rebuild confidence.

2) **Stabilize automation**
   - Move Claude billing/API usage behind:
     - budget alerts + auto-top-up (if possible)
     - fallback path that marks non-critical jobs as skipped (not failed) to preserve merge flow.

3) **Create a single canonical “Token Value Accrual & Cloud Revenue” doc**
   - Must explicitly define:
     - revenue vs yield
     - buyback mechanism (timing, % range if known, governance control)
     - future utility milestone tied to **Jeju**
   - Outcome metric: reduce repetitive Discord tokenomics questions week-over-week.

4) **Improve Cloud knowledge ingestion UX (even before raising limits)**
   - Add upload UI indicators: file size remaining, recommended preprocessing (markdown chunking), and an ingestion status page.
   - Provide an official “300 pages” playbook: split strategy + folder conventions.

5) **Onboarding consistency fix pack**
   - Pin “supported versions” (Bun/Node/NPM) + update monorepo docs.
   - Add CLI preflight check that warns about mismatched runtimes before install/build.

6) **Expectation management for custom model hosting**
   - If “next week” is firm: publish a minimal spec now (supported providers, self-hosted endpoints, auth).
   - If not firm: reframe as “phased rollout” with dates for alpha/beta to prevent credibility loss ahead of January ramp and EthDenver positioning.