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

FieldValue
Rolefounder
Decision authoritystrategic — all plans, agent SOUL edits, production deploys, API key rotation
Activeyes
Communication channeliMessage (Mac mini, +16194033648) + Discord (all ops / aurora-build channels)
HubSpot owneryes (primary account owner)
Discord userhenry / Henry Hill
Escalates ton/a — final decision authority
Agent handles domainAurora (iMessage routing) + Solara (strategic governance)
ClusterPersonnel 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:

  1. Message arrives at port 18802 (imessage-webhook handler)
  2. Aurora reads the chat GUID, matches allowlist, identifies context (business vs personal)
  3. If business (RERI / acquisitions / ops): Aurora responds with terse team-leader voice; escalates to appropriate agent if out of scope
  4. If personal (family chat with Stephania): warm/emoji tone per project_stephania_aurora_chat_wired §1.1.5
  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

ChannelHandle / NumberPrimary use
iMessage+16194033648Personal + ops coordination; routed via Aurora on Mac mini
DiscordHenry HillAll agent channels — approvals, escalation, plan review
Direct (local)VPS terminal / VS CodeCode 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

  1. 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
  2. Second Max Plan approved (2026-05-01)teamsteph@betterfiles.com Max plan on :18903 via anthropic-max-router serving 40 agents; henryRERI stays on interactive VSCode. Plan: project_second_max_plan_decision
  3. 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

Governance hubs

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)

Plans governing founder authority

Feedback rules

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 impaired since 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