Henry OS

AI Chief of Staff that autonomously finds leads, manages communications, schedules tasks, delivers software, and provides daily business briefings.

Security Scan

The skill's claims (deep access to inboxes, Apple apps, and web platforms) are not aligned with the minimal metadata and there are high-risk install instructions (curl | bash) pointing to an unknown domain — proceed only after more verification.

Purpose & Capability

The SKILL.md claims broad, continuous access to inboxes, Mail.app/iMessage, calendars, Upwork/LinkedIn/Reddit, and will run persistent sub-agents — yet the registry entry lists no required env vars, no config paths, and no included binaries or code. skill.json references an external installer and an entry file (install.sh) that is not included in the package. These are inconsistent: a skill that truly integrates with these services would require explicit credentials, OS-level permission steps, or included installers.

Instruction Scope

The runtime instructions tell the user to run a one-line installer (curl -fsSL https://henryos.ai/install.sh | bash) which will install a 24/7 agent that monitors private data sources (email, messages, calendars) and scrapes external services. The SKILL.md grants the agent discretion to continuously monitor multiple sources but does not document how credentials/consent are obtained, what exact data is collected, or where data is sent. Encouraging execution of remote scripts to gain broad local/data access is a major scope and privacy concern.

Install Mechanism

The SKILL.md and skill.json both direct installation by piping a script from henryos.ai to bash — this is a high-risk install pattern. The domain is not a well-known release host in the metadata, the archive/script is not included in the package for review, and there is no install manifest in the registry to verify what will be written to disk or what services will be created.

Credentials

Despite claiming integrations with many third-party platforms and local apps, the skill requests no environment variables, no API tokens, and no config paths. That is disproportionate: legitimate integrations typically require explicit API keys, OAuth flows, or documented local-permission steps. The absence of declared credentials or permission semantics is suspicious and prevents proper security assessment.

Persistence & Privilege

The skill intends to install a persistent 'Henry Master Agent' and multiple sub-agents that run 24/7 and expose a local web dashboard. 'always' is not set in the registry (so the platform won't force-enable it), but the install would create persistent background processes on the host. This persistent presence is reasonable for the claimed purpose only if the install artefacts and permission model are auditable — which are not provided here.

What to consider before installing

Do not run the one-line installer (curl | bash) or grant broad system permissions until you can verify the source and inspect the install script. Ask the publisher for: (1) the full install.sh contents or a link to a tagged GitHub release you can review, (2) a clear list of required OS permissions and exactly what data is read/sent and to which endpoints, (3) a privacy/security policy and code signatures or checksums for released artifacts, and (4) explicit instructions for OAuth or API keys for third-party services. If you must try it, run it in an isolated VM or disposable machine, review the installer script line-by-line, and avoid providing access to sensitive inboxes/contacts until you confirm what is being installed and where data flows.

Henry OS — AI Chief of Staff Framework

Stop being the bottleneck in your own business.

Henry OS transforms OpenClaw into a fully autonomous business partner that runs 24/7 — finding opportunities, managing your pipeline, handling communications, and executing tasks while you focus on high-leverage work.

---

What Henry OS Does

🎯 Proactive Intelligence

Henry doesn't wait for instructions. He continuously monitors your inboxes, calendars, job boards, and social channels — surfacing opportunities and flagging what needs attention before you even know it exists.

💰 Revenue Pipeline Automation

• Opportunity Hunting: Scans Upwork, LinkedIn, Reddit, Indie Hackers, and niche communities for freelance/consulting leads matching your skills
• Lead Scoring: Rates opportunities by fit, effort, and revenue potential
• Outreach Prep: Drafts personalized proposals and messages ready for your approval
• Pipeline Management: Tracks every lead from discovery to close in Mission Control

📧 Communications Management

• Inbox Triage: Reads and categorizes emails by urgency (Action Needed / FYI / Noise)
• Draft Replies: Prepares responses to common requests, waiting for your one-tap approval
• Follow-up Automation: Never drops the ball on pending conversations
• Cross-Channel: Works across Mail.app, Gmail, iMessage, SMS, WhatsApp

📅 Calendar & Life Ops

• Daily Briefings: Morning agenda with priorities, meetings, and action items
• Smart Scheduling: Blocks focus time around deep work, prepares for meetings
• Reminders & Follow-ups: Automatically creates follow-up tasks after calls and events
• Conflict Detection: Flags scheduling issues before they become problems

🛠️ Software Delivery

• Client Projects: End-to-end delivery from requirements to deployment
• Code Generation: Writes production-quality code in any language
• Testing & QA: Validates work before presenting it to you
• Documentation: Auto-generates READMEs and handoff docs

📺 YouTube Intelligence

• Content Processing: Watches videos, extracts actionable insights
• Implementation: Automatically applies relevant techniques to your active projects
• Knowledge Base: Builds a searchable library of insights from every video you consume

---

Installation

One-Line Install

curl -fsSL https://henryos.ai/install.sh | bash

That's it. Henry OS installs in under 5 minutes and starts working immediately.

What Gets Installed

• Mission Control Dashboard — Web UI at http://localhost:3333 for monitoring and management
• Henry Master Agent — Your autonomous operator running 24/7
• Specialist Sub-Agents — Scout, Builder, Writer, Analyst, Operator, Hunter, Watcher
• Native Apple Integrations — Mail.app, Messages.app, Calendar.app, Contacts.app
• Tooling Stack — All required CLIs and dependencies auto-configured

---

Quick Start

1. Launch Mission Control

After installation, open your browser to:

http://localhost:3333

This is your command center. From here you can:

• View system health and agent status
• Manage your task board (Kanban: Backlog / In Progress / Blocked / Done)
• Browse memory and learning loops
• Monitor your opportunity pipeline
• Chat directly with Henry

2. Configure Your Profile

Henry needs to know about you to be effective:

# Set your core details
henry config set name "Your Name"
henry config set email "[e-mail]"
henry config set timezone "Australia/Brisbane"

# Define your skills (comma-separated)
henry config set skills "web development,react,node,python,automation"

# Set your target rates
henry config set hourly_rate 150
henry config set project_min 2000

3. Connect Your Accounts

Henry works with your existing Apple apps — no new logins needed:

# Grant permissions (macOS will prompt)
henry setup permissions

# Verify connections
henry status

4. Start Your First Hunt

# Launch the opportunity hunter
henry spawn hunter

# Check what's been found
henry pipeline

5. Review Your Morning Brief

Every morning, Henry prepares a briefing:

• Overnight activity summary
• Today's calendar and priorities
• Top 3 opportunities to pursue
• Tasks requiring approval
• System health check

---

Configuration Options

Core Settings

Setting	Description	Default
name	Your name	System username
email	Primary email	—
timezone	Your timezone	System timezone
skills	Comma-separated skill list	—
hourly_rate	Target hourly rate	100
project_min	Minimum project value	1000

Opportunity Hunting

Setting	Description	Default
hunt.enabled	Enable automated hunting	true
hunt.frequency	Check interval (minutes)	60
hunt.sources	Sources to monitor	upwork,linkedin,reddit
hunt.min_budget	Minimum budget threshold	500

Communications

Setting	Description	Default
email.triage	Auto-triage emails	true
email.drafts	Draft replies for approval	true
messages.monitor	Monitor iMessage/SMS	true
approval.required	Require approval for sends	true

Notifications

Setting	Description	Default
notify.channel	Primary notification channel	imessage
notify.quiet_hours	Don't notify during these hours	23:00-08:00
notify.urgent_only	Only notify for urgent items	false

Advanced

# View all settings
henry config list

# Edit config file directly
henry config edit

# Reset to defaults
henry config reset

---

Mission Control Dashboard

The dashboard has six main sections:

1. System Health Bar

• Token usage and costs
• Uptime and agent count
• Current status indicators

2. Task Board

Live Kanban with four columns:

• Backlog — Queued work
• In Progress — Active tasks
• Blocked — Waiting on external input
• Done — Completed work

3. Memory Viewer

Three tabs:

• Episodic — Daily activity logs
• Semantic — Knowledge base (contacts, tools, learnings)
• Procedural — Playbooks and runbooks

4. Learning Loop Log

• Last 10 learning cycles
• Provisional lessons
• Promote/reject actions

5. Cron Monitor

• Scheduled tasks
• Run history
• Manual triggers

6. Approval Gate Banner

Amber full-width alert for anything needing your sign-off.

---

Sub-Agents

Henry spawns specialized agents for different work types:

Agent	Role	When Spawned
Henry-Scout	Research, scraping, monitoring	Intelligence gathering
Henry-Builder	Coding, automation, infrastructure	Software projects
Henry-Writer	Content, proposals, docs	Communications
Henry-Analyst	Data synthesis, scoring, modeling	Analysis tasks
Henry-Operator	Scheduling, inbox, reminders	Life ops
Henry-Hunter	Opportunity ID, lead research	Revenue pipeline
Henry-Watcher	YouTube monitoring, media intel	Content processing

Spawn agents manually:

henry spawn hunter    # Start hunting
henry spawn builder   # Begin a build task
henry spawn analyst   # Analyze data

Check agent status:

henry agents

---

Troubleshooting

Installation Issues

Problem: Install script fails with permission error

# Solution: Run with explicit bash
bash -c "$(curl -fsSL https://henryos.ai/install.sh)"

Problem: Node version too old

# Solution: Update Node.js first
brew install node  # macOS
# or
nvm install 20

Connection Issues

Problem: Mission Control won't load

# Check if server is running
henry status

# Restart Mission Control
henry restart dashboard

# Check logs
henry logs

Problem: Apple app permissions denied

# Re-run permission setup
henry setup permissions

# Or manually grant in:
# System Preferences → Security & Privacy → Privacy → Automation

Agent Issues

Problem: Agent won't spawn

# Check resource limits
henry status --resources

# View agent logs
henry logs --agent henry-hunter-01

Problem: Agent stuck/running too long

# List running agents
henry agents

# Terminate specific agent
henry terminate henry-hunter-01

# Kill all agents and restart
henry reset agents

Opportunity Hunting

Problem: No opportunities found

• Verify your skills are configured: henry config get skills
• Check hunt sources: henry config get hunt.sources
• Review minimum budget threshold
• Try manual search: henry hunt --manual "react developer"

Problem: Too many low-quality leads

• Increase minimum budget: henry config set hunt.min_budget 2000
• Refine skill keywords to be more specific
• Add negative keywords to exclude

Email/Communications

Problem: Not seeing emails in Mission Control

• Verify Mail.app has messages in inbox
• Check permissions: henry setup permissions
• Force sync: henry sync email

Problem: Drafts not being created

• Ensure email.drafts is enabled: henry config get email.drafts
• Check approval settings aren't blocking

Getting Help

# View all commands
henry --help

# Get help for specific command
henry hunt --help

# Check system diagnostics
henry doctor

# Export logs for support
henry logs --export

Community & Support

• Documentation: https://docs.henryos.ai
• GitHub Issues: https://github.com/shannonlinnan/henry-os/issues
• Discord: https://discord.gg/henryos
• Email: [e-mail]

---

Privacy & Security

Henry OS is designed with privacy-first principles:

• Local-First: Your data stays on your machine
• No Cloud Dependencies: Works entirely offline after installation
• Apple Native: Uses macOS native APIs, no third-party bridges
• Transparent: All code is open source and auditable
• Approval Gates: Sensitive actions require your explicit approval

---

Roadmap

• Core framework and Mission Control
• Native Apple integrations
• Opportunity hunting pipeline
• YouTube intelligence
• Windows/Linux support
• Browser extension for web clipping
• Mobile companion app
• Team/enterprise multi-agent mode
• Industry-specific agent packs

---

License

MIT License — see LICENSE for details.

---

Built with OpenClaw | Made by Shannon Linnan | ** henryos.ai **