Henry Hill — Founder
The anchor hub for all founder-level context. Henry Hill is the sole decision authority for OpenClaw: he approves all strategic plans, ratifies architecture decisions, controls agent identity changes, and owns all production credentials. Read this hub first when a task requires founder sign-off, escalation routing, or understanding communication preferences.
Quick reference
| Field | Value |
|---|---|
| Role | founder |
| Decision authority | strategic — all plans, agent SOUL edits, production deploys, API key rotation |
| Active | yes |
| Communication channel | iMessage (Mac mini, +16194033648) + Discord (all ops / aurora-build channels) |
| HubSpot owner | yes (primary account owner) |
| Discord user | henry / Henry Hill |
| Escalates to | n/a — final decision authority |
| Agent handles domain | Aurora (iMessage routing) + Solara (strategic governance) |
| Cluster | Personnel anchor: regional-acq-team, special-agents, stephania-historical |
Responsibilities
- Strategic direction — ratifies all multi-phase plans, OSIL blockers (B1–B16+), agent roadmap
- Agent identity governance — sole authority to approve SOUL.md edits; skill additions are additive and don’t require his approval unless they alter agent voice or capabilities fundamentally
- Credential stewardship — owns all platform API keys via 1Password vault Aurora; delegates rotation to Claude but retains vault admin
- Action-gate authority — all side-effect actions (file writes, service restarts, API calls with side effects) require explicit Henry approval unless session-level authorization was granted; see action-gate
- Plan Q&A — every open blocker question (Q1…Qn) must carry a recommendation before reaching Henry; he decides, doesn’t analyze
- Friction reduction — receives nightly friction reports; P0 items surface to Discord ops by 06:00 AM PST
Agent handoff pattern
When Henry sends an iMessage, Aurora (via BlueBubbles + imessage-responder skill) routes inbound messages:
- Message arrives at port 18802 (imessage-webhook handler)
- Aurora reads the chat GUID, matches allowlist, identifies context (business vs personal)
- If business (RERI / acquisitions / ops): Aurora responds with terse team-leader voice; escalates to appropriate agent if out of scope
- If personal (family chat with Stephania): warm/emoji tone per project_stephania_aurora_chat_wired §1.1.5
- Discord DMs from Henry land in relevant agent channels; Aurora monitors aurora-build
Action gate applies: Aurora NEVER sends outbound SMS/iMessage on Henry’s behalf without explicit session-level authorization per action-gate and feedback_action_gate_violation_repeated.
Communication channels
| Channel | Handle / Number | Primary use |
|---|---|---|
| iMessage | +16194033648 | Personal + ops coordination; routed via Aurora on Mac mini |
| Discord | Henry Hill | All agent channels — approvals, escalation, plan review |
| Direct (local) | VPS terminal / VS Code | Code review, log inspection, service control |
Mac mini context: Henry’s primary workstation. The BlueBubbles server runs here. iMessage → webhook → Aurora pipeline depends on this machine being online. Tailscale (openclaw-mac-ultra) provides connectivity.
VS Code workspace: /home/opsadmin/.openclaw/openclaw.code-workspace — multi-root, OC: prefixed tasks.
Recent decisions
- OSIL B1–B12 + B16 ratified (2026-05-03) — Approved the OpenClaw Self-Improvement Layer blockers including: B4=signoff first 4 weeks then auto-merge; B5=keep Hermes P2; B7=cloud free tier. Plan: openclaw-self-improvement-layer-2026-05-03
- Second Max Plan approved (2026-05-01) —
teamsteph@betterfiles.comMax plan on:18903viaanthropic-max-routerserving 40 agents;henryRERIstays on interactive VSCode. Plan: project_second_max_plan_decision - Wave 1/2 wikilink optimization ratified (2026-05-03) — Q1–Q7 visual mapping ratifications with 9 sharpenings; 50 hubs shipped (Wikilinks 308 → 7,142). Plan: openclaw-vault-wikilink-optimization-2026-05-03
Governance gates that apply to all interactions with Henry
- G-BLOCKER-SURFACING — unresolved plan blockers MUST mirror to SESSION-AUDIT.md NEXT ACTIONS. g-gates-network
- G-NO-PLAINTEXT-CREDS — no credential in any response or file. feedback_no_plaintext_creds
- G-PLAN-INDEX-REQUIRED — new plans require matching project_*.md + MEMORY.md index entry. feedback_plan_index_required
- Action gate (2026-03-31) — reads/searches/status queries → act directly; writes/deletes/restarts/API calls → state and wait for confirmation. action-gate
Cross-links
Governance hubs
- decision-log — ratification history
- plan-governance — plan creation + amendment standards
- action-gate — side-effect approval protocol
- g-gates-network — all 8 cascade-failure gates
- sources-first — sources protocol Henry enforces
- blockers-first — blocker surfacing Henry expects
- memory-rule-clusters — 92 feedback rules indexed
Agents that report to Henry
- _summary — Chief AI Operations, primary iMessage router
- _summary — Strategic governance, escalation chain
- _summary — InvestorLift + infra, strategic asset
Resources
- henry-preferences — HENRY-PREFERENCES.md (formatting, question style, review standards)
Related team hubs
- regional-acq-team — David, Ryan, hhill field team
- special-agents — Aurora, Atlas, Solara identities
- stephania-historical — historical context for Stephania iMessage routing
Plans governing founder authority
- openclaw-fragmentation-fix-2026-05-01 — governance gates §A1
- openclaw-self-improvement-layer-2026-05-03 — OSIL master plan, Henry the ratifying authority
Feedback rules
- feedback_action_gate_violation_repeated — action gate enforcement
- feedback_question_discipline — how to bring questions to Henry
- feedback_options_listed_vertically — question format standard
- feedback_no_assumptions — verification before claiming X exists
Open issues / TODOs
- Root SSH key rotation P0 pending (used in 2026-05-03 session, in master.env) — Henry must rotate via AWS Console
- B13a / B14a / B15 OSIL blockers pending Henry decision
- Mac Ultra
impairedsince 2026-05-02 22:15 UTC — resume via AWS Console reboot
Recent activity
- 2026-05-04: hub created (W3-S2)
- 2026-05-03: OSIL B1–B16 ratifications; wikilink Wave 1/2 Q&A ratifications
- 2026-05-01: Second Max Plan approval; VPS cascade remediation F1–F4 authorization