### Install KyberBot

Source: https://github.com/kybernesisai/kyberbot/blob/main/README.md

Clone the repository, install dependencies, build the project, and link the CLI package. This is the initial setup for KyberBot.

```bash
git clone https://github.com/KybernesisAI/kyberbot.git
cd kyberbot && npm install && npm run build
cd packages/cli && npm link && cd ../..
```

--------------------------------

### Example Output of 'kyberbot skill list'

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/skills.md

This is an example of the output you can expect when listing installed skills.

```bash
$ kyberbot skill list

  Skills

  NAME              VERSION  DESCRIPTION
  morning-briefing  1.0.0    Compile a morning briefing from all sources
  weekly-report     1.0.0    Generate a weekly summary of completed work
  water-tracker     1.0.0    Track daily water intake
```

--------------------------------

### Install and Initialize KyberBot Agent

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/why-kyberbot.md

Installs KyberBot globally, links the CLI package, and then initializes a new agent in a specified directory. This is the initial setup for creating a personal AI agent.

```bash
git clone https://github.com/KybernesisAI/kyberbot.git
cd kyberbot
npm install && npm run build
cd packages/cli && npm link && cd ../..

mkdir ~/my-agent && cd ~/my-agent
kyberbot onboard

kyberbot              # Start services (leave running)
claude                # New terminal — talk to your agent
```

--------------------------------

### Install KyberBot CLI

Source: https://github.com/kybernesisai/kyberbot/blob/main/README.md

Clone the repository, install dependencies, build the project, and link the CLI globally. This is a one-time setup for the KyberBot application itself.

```bash
git clone https://github.com/KybernesisAI/kyberbot.git
cd kyberbot
npm install
npm run build
cd packages/cli && npm link && cd ../..
```

--------------------------------

### Run Skill Setup

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/skills.md

Execute the setup script for a skill if it has specific setup requirements.

```bash
# Run skill setup (if the skill has setup requirements)
kyberbot skill setup <name>
```

--------------------------------

### Example: Recall Project Status

Source: https://github.com/kybernesisai/kyberbot/blob/main/template/skills/recall/SKILL.md

To get the status of a project, use 'recall' with the project name. This example demonstrates recalling information about the 'dashboard redesign'. It can be combined with timeline search for more context.

```bash
kyberbot recall "dashboard redesign"
```

```bash
kyberbot timeline --search "dashboard"
```

--------------------------------

### Manage Skills

Source: https://context7.com/kybernesisai/kyberbot/llms.txt

Commands for managing skills, which are reusable capabilities. Use 'list' (with --json option) to see installed skills, 'create' to make new ones (with options for description, environment variables, and setup), 'info' to show details, 'remove' to delete, 'setup' to run a skill's setup script, and 'rebuild' to update CLAUDE.md.

```bash
kyberbot skill list
```

```bash
kyberbot skill list --json
```

```bash
kyberbot skill create my-skill
```

```bash
kyberbot skill create weekly-report --description "Generate weekly summary"
```

```bash
kyberbot skill create strava-sync --env STRAVA_TOKEN --setup
```

```bash
kyberbot skill info my-skill
```

```bash
kyberbot skill remove old-skill
```

```bash
kyberbot skill setup strava-sync
```

```bash
kyberbot skill rebuild
```

--------------------------------

### Example Agent Conversation Prompts

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/getting-started.md

Provides example prompts to interact with your KyberBot agent. These demonstrate how to ask for information, provide new data, and request task creation.

```bash
> Hey Atlas, what do you know about me?
```

```bash
> Remember that my product launch deadline is June 15th.
```

```bash
> What's on my schedule today?
```

```bash
> Create a skill for tracking my running mileage.
```

--------------------------------

### Install Node.js, Docker, Claude Code, and pm2

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/vps-deployment.md

Installs necessary dependencies for KyberBot on a Linux VPS. Ensure Node.js is installed via nvm, Docker is set up, Claude Code is globally installed via npm, and pm2 is used for process management.

```bash
# Node.js (via nvm)
curl -o- https://raw.githubusercontent.com/nvm-sh/nvm/v0.40.1/install.sh | bash
source ~/.bashrc
nvm install 22

# Docker
curl -fsSL https://get.docker.com | sh
sudo usermod -aG docker $USER
newgrp docker

# Claude Code
npm install -g @anthropic-ai/claude-code

# pm2 (process manager)
npm install -g pm2
```

--------------------------------

### SKILL.md Format Example

Source: https://context7.com/kybernesisai/kyberbot/llms.txt

An example of the SKILL.md format, which uses YAML frontmatter for metadata and markdown for instructions and implementation details. This example defines a water tracker skill.

```markdown
---
name: water-tracker
description: Track daily water intake and show totals
version: 1.0.0
requires_env: []
has_setup: false
---

# Water Tracker

## What This Does

Tracks glasses of water consumed throughout the day and provides daily/weekly summaries.

## How to Use

- "log water"
- "I drank a glass of water"
- "water intake today"
- "how much water did I drink"

## Implementation

1. When the user logs water, extract the count (default: 1 glass)
2. Store to brain with tag "water-intake" and today's date
3. When queried for totals, search brain for water-intake tags
4. Calculate and display daily total, weekly average
5. Compare against 8 glasses/day goal from USER.md

```

--------------------------------

### Create and Start a New KyberBot Agent

Source: https://context7.com/kybernesisai/kyberbot/llms.txt

Create a new agent instance in a directory and start all its background services. You can then interact with the agent from a separate terminal.

```bash
mkdir ~/my-agent
cd ~/my-agent
kyberbot onboard
kyberbot
cd ~/my-agent
claude
```

--------------------------------

### Install Project Dependencies

Source: https://github.com/kybernesisai/kyberbot/blob/main/CONTRIBUTING.md

Install all project dependencies for the monorepo using npm.

```bash
npm install
```

--------------------------------

### Create Skill with Options

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/skills.md

When creating a skill, you can specify a description and environment variables or setup requirements.

```bash
kyberbot skill create my-skill -d "Track weekly running mileage"
kyberbot skill create my-skill --env STRAVA_TOKEN --setup
```

--------------------------------

### Manage Sub-Agents

Source: https://context7.com/kybernesisai/kyberbot/llms.txt

Commands for managing sub-agents, which are specialized AI instances. Use 'list' (with --json option) to see installed agents, 'create' to make new ones (with options for description, role, model, and max turns), 'info' to show details, and 'spawn' to start an agent with a specific task.

```bash
kyberbot agent list
```

```bash
kyberbot agent list --json
```

```bash
kyberbot agent create researcher
```

```bash
kyberbot agent create code-reviewer --description "Review PRs" --role "Senior engineer"
```

```bash
kyberbot agent create fast-helper --model haiku --max-turns 5
```

```bash
kyberbot agent info researcher
```

```bash
kyberbot agent spawn researcher "Find information about Kubernetes best practices"
```

--------------------------------

### WhatsApp Channel Setup

Source: https://context7.com/kybernesisai/kyberbot/llms.txt

Steps to configure WhatsApp messaging using QR code authentication.

```bash
# 1. Add WhatsApp channel
kyberbot channel add whatsapp

# 2. QR code appears in terminal
#    Open WhatsApp > Settings > Linked Devices > Link a Device
#    Scan the QR code

# 3. Terminal shows: WhatsApp connected, session saved

# 4. Chat with your agent via WhatsApp
```

--------------------------------

### Example: Search Kybernesis Cloud

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/kybernesis.md

Perform a specific search query against your Kybernesis Cloud workspace memory. Replace '...' with your actual search terms.

```bash
# Search for something specific
kyberbot kybernesis query "What do I know about project pricing?"
```

--------------------------------

### Core Commands

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/cli-reference.md

Commands for starting services, setting up agents, updating the CLI, and checking service status.

```APIDOC
## Core Commands

### `kyberbot run`

Start all services.

**Options:** `--no-channels`, `--no-sleep`, `--no-heartbeat`, `-v/--verbose`

### `kyberbot onboard`

Set up a new agent.

### `kyberbot update`

Update CLI and refresh templates.

**Options:** `--check` (dry-run), `--templates` (skip CLI source update)

### `kyberbot status`

Show service health dashboard.

**Options:** `--json`
```

--------------------------------

### List All Installed Skills

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/skills.md

Use the 'kyberbot skill list' command to display all skills currently installed in the agent.

```bash
# List all installed skills
kyberbot skill list
```

--------------------------------

### Start KyberBot

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/channels.md

This command starts the KyberBot server. If a new Telegram bot is configured, it enters verification mode.

```bash
kyberbot
```

--------------------------------

### User Information Configuration

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/getting-started.md

Example prompts for providing personal details to the agent, populating USER.md.

```bash
Your name:
> Alex

Timezone (detected: America/New_York):
> America/New_York

Location (optional):
> New York

Tell your agent something about yourself (optional):
> I'm a software engineer working on fintech products.
```

--------------------------------

### KyberBot Onboarding Wizard Command

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/getting-started.md

Re-runs the initial onboarding wizard for setting up a new agent. This is useful for reconfiguring or starting fresh with an agent instance.

```bash
kyberbot onboard
```

--------------------------------

### Example Telegram Chat Interaction

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/channels.md

Illustrates a typical conversation with the KyberBot agent via Telegram after successful setup and verification.

```text
You: Hey, what's on my schedule today?
Bot: Here's your schedule for today...
```

--------------------------------

### Verify Prerequisites with Node.js, Docker, and Claude

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/getting-started.md

Run these commands to ensure Node.js, Docker, and Claude Code are installed and accessible.

```bash
node --version    # Should be 18.0.0 or higher
docker --version  # Should show Docker version
claude --version  # Should show Claude Code version
```

--------------------------------

### List Installed KyberBot Skills

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/getting-started.md

Shows a list of all skills that are currently installed and available for the KyberBot agent. This helps in understanding the agent's capabilities.

```bash
kyberbot skill list
```

--------------------------------

### Specific Skill Implementation Example

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/skills.md

Illustrates a detailed and specific implementation for a skill, contrasting it with a vague one. Use this pattern to clearly define skill actions and data extraction.

```markdown
## Implementation
Help the user track expenses.
```

```markdown
## Implementation
1. When the user logs an expense, extract: amount, category, date, description
2. Categories: food, transport, housing, health, entertainment, other
3. Store in brain with tag "expense" and the category as a second tag
4. When queried for totals, group by category and show amounts
5. Compare against monthly budget if defined in USER.md
```

--------------------------------

### Telegram Channel Setup

Source: https://context7.com/kybernesisai/kyberbot/llms.txt

Steps to configure Telegram messaging, including bot creation and verification.

```bash
# 1. Create bot with @BotFather in Telegram
#    Send /newbot, follow prompts, get token

# 2. Add channel
kyberbot channel add telegram
# Enter token when prompted

# 3. Start KyberBot
kyberbot
# Console shows: Send /start ABCDEF to your Telegram bot

# 4. In Telegram, send to your bot:
/start ABCDEF

# 5. Bot responds: Connected! You are now the verified owner.

# 6. Now chat normally with your agent via Telegram
```

--------------------------------

### Example WhatsApp Chat Interaction

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/channels.md

Demonstrates a sample interaction with the KyberBot agent through WhatsApp.

```text
You: What did I accomplish yesterday?
Bot: Based on your timeline, yesterday you...
```

--------------------------------

### Example of a new research findings note

Source: https://github.com/kybernesisai/kyberbot/blob/main/template/skills/brain-note/SKILL.md

This example demonstrates creating a new markdown file for research findings in the brain directory.

```markdown
brain/kubernetes-migration.md (new file):

# Kubernetes Migration Research — 2025-02-23

## Current State
Running on bare EC2 instances with manual deploys...

## Findings
...

```

--------------------------------

### Agent Identity Configuration

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/getting-started.md

Example prompts for setting the agent's name, description, and personality.

```bash
What should your AI agent be called?
> Atlas

One-line description of your agent:
> My personal AI agent

How would you like to define its personality? (SOUL.md)
> Guided template (recommended)
```

--------------------------------

### Example: Recall Past Decision

Source: https://github.com/kybernesisai/kyberbot/blob/main/template/skills/recall/SKILL.md

When a user asks about a past decision, use 'recall' to find information about the relevant entity and 'search' for broader context. This example queries about 'PostgreSQL'.

```bash
kyberbot recall "PostgreSQL"
```

```bash
kyberbot search "PostgreSQL decision"
```

--------------------------------

### Example: Recall Person

Source: https://github.com/kybernesisai/kyberbot/blob/main/template/skills/recall/SKILL.md

When a user mentions a person, use the 'recall' command with their name to retrieve relevant information. This example shows recalling information about 'Jake'.

```bash
kyberbot recall "Jake"
```

--------------------------------

### Example: List Recent Memories with Pagination

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/kybernesis.md

Retrieve a paginated list of recent memories from your Kybernesis Cloud workspace. Adjust the --limit and --offset parameters as needed.

```bash
# List recent memories with pagination
kyberbot kybernesis list --limit 20 --offset 0
```

--------------------------------

### Start KyberBot for WhatsApp QR Code Scan

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/vps-deployment.md

Starts KyberBot in the foreground to display a QR code for WhatsApp authentication. This is necessary for the initial connection of WhatsApp to KyberBot.

```bash
# Start in foreground for QR scan
kyberbot run

# Scan the QR code from your phone
# After connecting, stop and switch to pm2
pm2 start "kyberbot run" --name kyberbot --cwd ~/my-agent
```

--------------------------------

### Example of an appended architecture decision note

Source: https://github.com/kybernesisai/kyberbot/blob/main/template/skills/brain-note/SKILL.md

This example shows how to append an architecture decision to an existing markdown file in the brain directory.

```markdown
brain/architecture-decisions.md (append):

## Frontend Framework Choice — 2025-02-23

**Context**: Needed to choose a framework for the new dashboard
**Decision**: Next.js with App Router
**Rationale**: SSR support, team familiarity, strong ecosystem
**Alternatives considered**: Remix (less mature), SvelteKit (team unfamiliar)
**Implications**: Lock into React ecosystem, need to learn App Router patterns
```

--------------------------------

### Claude Connection Method

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/getting-started.md

Example prompt for selecting the connection method to Claude for background operations.

```bash
How would you like to connect to Claude?
> Agent SDK (recommended) -- works with your Claude Code subscription
```

--------------------------------

### SKILL.md Format Example

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/skills.md

A SKILL.md file includes YAML frontmatter for metadata and natural language instructions for the agent.

```markdown
---
name: skill-name
description: One-line description of what this skill does
version: 1.0.0
requires_env: []
has_setup: false
---

# Skill Name

## What This Does

A brief explanation of what this skill accomplishes and when it should be used.

## How to Use

- "trigger phrase one"
- "trigger phrase two"
- "trigger phrase three"

## Implementation

Step-by-step instructions the agent follows when executing this skill.

1. First step
2. Second step
3. Third step
```

--------------------------------

### Messaging Channels Connection

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/getting-started.md

Example prompt for connecting messaging channels like Telegram or WhatsApp.

```bash
Connect messaging channels? (Telegram / WhatsApp)
> No
```

--------------------------------

### Start KyberBot Services

Source: https://github.com/kybernesisai/kyberbot/blob/main/README.md

Start the KyberBot services in one terminal and interact with your agent via the 'claude' command in a second terminal. Ensure KyberBot services are left running.

```bash
kyberbot          # Start services (leave running)
# Open a second terminal:
claude            # Talk to your agent
```

--------------------------------

### Cloud Sync / Kybernesis Connection

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/getting-started.md

Example prompt for enabling or disabling cloud memory sync via Kybernesis.

```bash
Enable cloud memory sync via Kybernesis? (optional)
> No
```

--------------------------------

### ESM Import Example

Source: https://github.com/kybernesisai/kyberbot/blob/main/CONTRIBUTING.md

Demonstrates the correct way to import local modules in an ES module environment, requiring `.js` extensions.

```typescript
// Correct
import { getConfig } from './config.js';
import { SearchResult } from './types.js';

// Incorrect
import { getConfig } from './config';
import { getConfig } from './config.ts';
```

--------------------------------

### Start Claude Code with Agent

Source: https://github.com/kybernesisai/kyberbot/blob/main/README.md

Open a second terminal, navigate to your agent's folder, and start Claude Code. It will automatically load your agent's personality and memory.

```bash
cd ~/my-agent
claude
```

--------------------------------

### Clone, Build, and Link KyberBot

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/vps-deployment.md

Clones the KyberBot repository, installs dependencies, builds the project, and links the CLI package for use. This prepares the KyberBot application for agent creation.

```bash
# Clone and build
git clone https://github.com/KybernesisAI/kyberbot.git
cd kyberbot
npm install && npm run build
cd packages/cli && npm link && cd ../..
```

--------------------------------

### Skill Trigger Phrases Example

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/skills.md

Provides examples of natural language trigger phrases for a skill. Include variations to improve the agent's ability to match user messages to the correct skill.

```markdown
## How to Use

- "log a run"
- "I went running"
- "running mileage"
- "how far did I run"
- "track my run"
```

--------------------------------

### USER.md Example: Week 1 Work Information

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/living-documents.md

Initial work-related information added to USER.md in Week 1.

```markdown
## Work

Software engineer. Building a SaaS product. Uses TypeScript and Next.js.

```

--------------------------------

### Start KyberBot with Specific Services Disabled

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/getting-started.md

Allows starting the KyberBot runtime while disabling specific services like the sleep agent, messaging channels, or the heartbeat scheduler. Useful for customizing the agent's operational footprint.

```bash
kyberbot --no-sleep
```

```bash
kyberbot --no-channels
```

```bash
kyberbot --no-heartbeat
```

--------------------------------

### Run KyberBot in Foreground for Verification

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/vps-deployment.md

Starts the KyberBot server in the foreground to verify that all services are running correctly. Press Ctrl+C to stop after confirmation.

```bash
cd ~/my-agent
kyberbot run
```

--------------------------------

### SOUL.md Example: Day 1 Communication Style

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/living-documents.md

An initial, broad communication style for an agent on Day 1.

```markdown
## Communication Style

Be helpful and clear. Ask clarifying questions when uncertain.

```

--------------------------------

### Example: Check Kybernesis Cloud Status

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/kybernesis.md

Verify the connectivity and reachability of the Kybernesis Cloud service from your KyberBot agent.

```bash
# Check if Kybernesis Cloud is reachable
kyberbot kybernesis status
```

--------------------------------

### Show KyberBot Skill Details

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/getting-started.md

Displays detailed information about a specific installed skill, including its purpose, parameters, and usage. Use this to understand how a particular skill functions.

```bash
kyberbot skill info my-skill
```

--------------------------------

### SOUL.md Example: Day 60 Communication Style

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/living-documents.md

A detailed and specific communication style for an agent on Day 60, reflecting learned preferences.

```markdown
## Communication Style

- Lead with the answer. Context comes after.
- Use bullet points for any list of 3+ items.
- Code examples over prose when explaining technical concepts.
- Never say "Great question!" or use filler phrases.
- When uncertain, say "I'm not sure" -- do not hedge or qualify.
- Morning briefings: max 10 bullet points, sorted by priority.
- Evening reviews: ask 3 reflective questions, then summarize.
- Default to action ("Here's the draft") over permission ("Would you like me to...").

```

--------------------------------

### USER.md Example: Week 8 Work Information

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/living-documents.md

Expanded work-related information in USER.md by Week 8, including role, projects, tech stack, team, and work patterns.

```markdown
## Work

### Current Role
Senior engineer at a fintech startup (Series A, 12 people).

### Active Projects
- **PayFlow** -- Payment processing dashboard. Next.js 15, Supabase, Stripe.
  Target launch: June 15. Current status: Beta testing with 3 pilot customers.
- **Internal Tools** -- Admin dashboard for ops team. Low priority.

### Tech Stack
TypeScript, Next.js 15, React 19, Supabase, Tailwind CSS, Vercel.
Prefers Cursor for IDE, Claude Code for CLI.

### Team
- Sarah (PM) -- Manages sprint planning. Prefers Slack over email.
- Marcus (design) -- Figma files in shared workspace.
- Priya (backend) -- Handles Stripe integration.

### Work Patterns
- Deep work: 9-12 AM (no meetings)
- Standup: 12:30 PM daily
- Sprint planning: Monday 2 PM
- Prefers async communication for non-urgent items

```

--------------------------------

### Start KyberBot with pm2 for Background Operation

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/vps-deployment.md

Manages KyberBot as a background process using pm2, ensuring it restarts on crashes and survives server reboots. This is the recommended method for 24/7 uptime.

```bash
# Start KyberBot as a managed process
pm2 start "kyberbot run" --name kyberbot --cwd ~/my-agent

# Save the process list so pm2 restores it on reboot
pm2 save

# Generate and install the systemd startup script
pm2 startup
# Run the command it prints (sudo env PATH=...)
```

--------------------------------

### Add WhatsApp Channel to KyberBot

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/channels.md

Initiates the WhatsApp Web authentication flow. This command starts the process to link your WhatsApp account to KyberBot.

```bash
kyberbot channel add whatsapp
```

--------------------------------

### Create and Onboard Agent

Source: https://github.com/kybernesisai/kyberbot/blob/main/README.md

Create a new directory for your agent and use the KyberBot CLI to onboard it. This prepares your agent for use.

```bash
mkdir ~/my-agent && cd ~/my-agent
kyberbot onboard
```

--------------------------------

### Manually Create Skill Directory and File

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/skills.md

Manually create the necessary directory and SKILL.md file for a new skill.

```bash
mkdir -p skills/my-skill
touch skills/my-skill/SKILL.md
```

--------------------------------

### Conventional Commits Example

Source: https://github.com/kybernesisai/kyberbot/blob/main/CONTRIBUTING.md

Examples of commit messages following the Conventional Commits specification.

```git
feat: add WhatsApp channel support
fix: prevent duplicate entity graph entries
docs: update brain architecture documentation
refactor: extract sleep cycle steps into separate modules
chore: update dependencies
```

--------------------------------

### Conventional Commits with Scope Example

Source: https://github.com/kybernesisai/kyberbot/blob/main/CONTRIBUTING.md

Examples of commit messages using a scope to specify the affected package.

```git
feat(cli): add timeline query command
fix(create-kyberbot): handle spaces in project path
```

--------------------------------

### Create Skill via CLI

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/skills.md

Use the KyberBot CLI to scaffold a new skill directory and its SKILL.md file.

```bash
kyberbot skill create my-skill
```

--------------------------------

### Create a New Agent

Source: https://github.com/kybernesisai/kyberbot/blob/main/README.md

Navigate to your desired agent directory and run the 'kyberbot onboard' command to initialize a new agent. This sets up personality files, memory databases, and Claude Code configuration.

```bash
mkdir ~/my-agent
cd ~/my-agent
kyberbot onboard
```

--------------------------------

### Create a New Skill

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/skills.md

Initiate the creation of a new skill using the CLI, providing the skill's name.

```bash
# Create a new skill
kyberbot skill create <name>
```

--------------------------------

### Start KyberBot with Specific Services Disabled

Source: https://context7.com/kybernesisai/kyberbot/llms.txt

Control which background services start with KyberBot. Use flags like --no-channels, --no-sleep, or --no-heartbeat to disable specific functionalities.

```bash
kyberbot --no-channels
```

```bash
kyberbot --no-sleep
```

```bash
kyberbot --no-heartbeat
```

--------------------------------

### Rebuild CLAUDE.md

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/skills.md

Update the CLAUDE.md file to reflect the current set of installed skills.

```bash
# Rebuild CLAUDE.md with current skills
kyberbot skill rebuild
```

--------------------------------

### Create a New KyberBot Agent

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/vps-deployment.md

Initializes a new KyberBot agent in a dedicated directory. This command sets up the necessary configuration and files for a new agent instance.

```bash
# Create your agent
mkdir ~/my-agent && cd ~/my-agent
kyberbot onboard
```

--------------------------------

### File Structure Reference

Source: https://context7.com/kybernesisai/kyberbot/llms.txt

Overview of the key files and directories within a KyberBot agent instance.

```APIDOC
## File Structure Reference

### Description
Key files and directories in a KyberBot agent instance.

```
my-agent/
├── SOUL.md                    # Agent personality (living document)
├── USER.md                    # User knowledge (living document)
├── HEARTBEAT.md               # Recurring tasks (living document)
├── identity.yaml              # Agent identity config
├── .env                       # Secrets (API keys, tokens)
├── heartbeat-state.json       # Scheduler state
├── docker-compose.yml         # Memory services
│
├── brain/                     # Markdown knowledge files
│   ├── projects/
│   ├── people/
│   └── decisions/
│
├── skills/                    # Installed skills
│   └── my-skill/
│       └── SKILL.md
│
├── data/                      # Runtime data (gitignored)
│   ├── chromadb/
│   ├── entity-graph.db
│   ├── timeline.db
│   └── sleep.db
│
├── .claude/
│   ├── CLAUDE.md
│   ├── settings.local.json
│   └── agents/
│
└── logs/                      # Application logs
    ├── heartbeat.log
    ├── sleep.log
    └── channels.log
```
```

--------------------------------

### Create a Feature Branch

Source: https://github.com/kybernesisai/kyberbot/blob/main/CONTRIBUTING.md

Create a new feature branch from the main branch to start making changes.

```bash
git checkout -b feat/my-feature
```

--------------------------------

### Create and Restore KyberBot Backups

Source: https://context7.com/kybernesisai/kyberbot/llms.txt

Commands for creating agent data backups and restoring them.

```bash
kyberbot backup
```

```bash
kyberbot backup --include brain,skills
```

--------------------------------

### KyberBot File Structure

Source: https://context7.com/kybernesisai/kyberbot/llms.txt

Illustrates the typical directory and file structure of a KyberBot agent instance, outlining key configuration, data, and knowledge files.

```plaintext
my-agent/
├── SOUL.md                    # Agent personality (living document)
├── USER.md                    # User knowledge (living document)
├── HEARTBEAT.md               # Recurring tasks (living document)
├── identity.yaml              # Agent identity config
├── .env                       # Secrets (API keys, tokens)
├── heartbeat-state.json       # Scheduler state
├── docker-compose.yml         # Memory services
│
├── brain/                     # Markdown knowledge files
│   ├── projects/              # Project-specific knowledge
│   ├── people/                # People profiles
│   └── decisions/             # Decision logs
│
├── skills/                    # Installed skills
│   └── my-skill/
│       └── SKILL.md
│
├── data/                      # Runtime data (gitignored)
│   ├── chromadb/              # Semantic search storage
│   ├── entity-graph.db        # Entity relationships
│   ├── timeline.db            # Temporal event log
│   └── sleep.db               # Sleep agent state
│
├── .claude/                   # Claude Code configuration
│   ├── CLAUDE.md              # Auto-generated instructions
│   ├── settings.local.json    # Permissions
│   └── agents/                # Sub-agent definitions
│
└── logs/                      # Application logs
    ├── heartbeat.log
    ├── sleep.log
    └── channels.log
```

--------------------------------

### Show Skill Details

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/skills.md

Retrieve detailed information about a specific skill using its name.

```bash
# Show details of a specific skill
kyberbot skill info <name>
```

--------------------------------

### Restart ChromaDB Docker Container

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/vps-deployment.md

Restarts the ChromaDB Docker container if it fails to start. This command ensures that the database service is running correctly for KyberBot.

```bash
cd ~/my-agent
docker compose up -d
kyberbot status
```

--------------------------------

### Run KyberBot Backup

Source: https://github.com/kybernesisai/kyberbot/blob/main/template/skills/backup/SKILL.md

Execute the full backup process, including checkpointing SQLite WAL files, syncing Claude Code memory, committing changes, and pushing to GitHub. This is the primary command for performing a backup.

```bash
kyberbot backup run
```

--------------------------------

### Telegram Bot Verification

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/channels.md

After starting KyberBot with a new Telegram bot, send this command to your bot in Telegram to verify ownership. Replace ABCDEF with the code displayed in the console.

```bash
/start ABCDEF
```

--------------------------------

### Enable Verbose Output for KyberBot

Source: https://context7.com/kybernesisai/kyberbot/llms.txt

Enable verbose logging for debugging purposes. Use the -v or --verbose flags to get detailed output from the KyberBot CLI.

```bash
kyberbot -v
```

```bash
kyberbot --verbose
```

--------------------------------

### Browse Cloud Workspace Memories

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/getting-started.md

Use this command to browse all available memories in your cloud workspace.

```bash
kyberbot kybernesis list
```

--------------------------------

### Link CLI Globally for Testing

Source: https://github.com/kybernesisai/kyberbot/blob/main/CONTRIBUTING.md

Link the KyberBot CLI globally to test it locally. After linking, you can use the 'kyberbot' command.

```bash
# Link the CLI globally for testing
cd packages/cli
npm link

# Now you can use the kyberbot command
kyberbot --help
```

--------------------------------

### Channel Interface Definition

Source: https://context7.com/kybernesisai/kyberbot/llms.txt

Defines the interface for custom messaging channels, including methods for starting, stopping, sending messages, checking connection status, and handling incoming messages.

```typescript
interface Channel {
  /** Unique channel identifier */
  readonly name: string;

  /** Initialize the channel (connect, authenticate) */
  start(): Promise<void>;

  /** Gracefully shut down the channel */
  stop(): Promise<void>;

  /** Send a message to a recipient */
  send(to: string, message: string): Promise<void>;

  /** Check if the channel is currently connected */
  isConnected(): boolean;

  /** Register a handler for incoming messages */
  onMessage(handler: (message: ChannelMessage) => Promise<void>): void;
}

interface ChannelMessage {
  id: string;
  channelType: string;  // 'telegram', 'whatsapp', etc.
  from: string;
  text: string;
  timestamp: Date;
  metadata?: Record<string, unknown>;
}
```

--------------------------------

### Weekly Planning Task Instructions

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/living-documents.md

Defines the instructions for the 'Weekly Planning' task, specifying the cadence, day, time, and content for weekly planning. This task is enabled by default.

```markdown
### Weekly Planning
- **cadence**: weekly
- **day**: monday
- **time**: 09:00
- **enabled**: true

#### Instructions

Compile a weekly planning session:
1. Review last week's accomplishments
2. Check project milestones and deadlines
3. Identify blockers
4. Propose this week's goals (max 5)
5. Flag any upcoming meetings that need prep
```

--------------------------------

### Adding a New Discord Channel

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/channels.md

Illustrates the file path for creating a new channel implementation, specifically for Discord. This involves implementing the Channel interface and registering it within the server index.

```bash
packages/cli/src/server/channels/discord.ts
```

--------------------------------

### Query Specific Entity in KyberBot

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/getting-started.md

Retrieves detailed information about a specific entity from the agent's memory. Use this to get context or details about a particular person, project, or concept.

```bash
kyberbot recall "John"
```

--------------------------------

### List All Tracked Entities

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/brain.md

Lists all entities currently being tracked by the KyberBot's entity graph. This command is useful for getting an overview of all known people, companies, projects, and topics.

```bash
# List all tracked entities
kyberbot recall
```

--------------------------------

### Run Type Checks

Source: https://github.com/kybernesisai/kyberbot/blob/main/CONTRIBUTING.md

Execute type checking for the entire project.

```bash
npm run typecheck
```

--------------------------------

### Heartbeat State Management

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/living-documents.md

Example JSON structure for `heartbeat-state.json`, which tracks the execution state of scheduled tasks, including last run time, next run time, status, and run counts.

```json
{
  "tasks": {
    "morning-briefing": {
      "lastRun": "2026-02-22T08:00:00Z",
      "nextRun": "2026-02-23T08:00:00Z",
      "lastStatus": "success",
      "runCount": 45,
      "consecutiveFailures": 0
    },
    "evening-review": {
      "lastRun": "2026-02-21T21:00:00Z",
      "nextRun": "2026-02-22T21:00:00Z",
      "lastStatus": "success",
      "runCount": 44,
      "consecutiveFailures": 0
    }
  },
  "timezone": "America/New_York",
  "version": 1
}
```

--------------------------------

### TypeScript Channel Interface Definition

Source: https://github.com/kybernesisai/kyberbot/blob/main/docs/channels.md

Defines the contract for a messaging channel, including methods for starting, stopping, sending messages, checking connection status, and handling incoming messages. Requires a ChannelMessage interface for message structure.

```typescript
interface Channel {
  /** Unique channel identifier */
  readonly name: string;

  /** Initialize the channel (connect, authenticate) */
  start(): Promise<void>;

  /** Gracefully shut down the channel */
  stop(): Promise<void>;

  /** Send a message to a recipient */
  send(to: string, message: string): Promise<void>;

  /** Check if the channel is currently connected */
  isConnected(): boolean;

  /** Register a handler for incoming messages */
  onMessage(handler: (message: ChannelMessage) => Promise<void>): void;
}

interface ChannelMessage {
  /** Unique message identifier */
  id: string;

  /** Channel type (e.g., 'telegram', 'whatsapp') */
  channelType: string;

  /** Sender identifier */
  from: string;

  /** Raw text content */
  text: string;

  /** Timestamp */
  timestamp: Date;

  /** Optional extra data */
  metadata?: Record<string, unknown>;
}
```