AI-powered enterprise execution by Kam Solutions. Turn strategy, requirements, and operations into structured action through intelligent agents, workflow automation, and business-aware execution.
KAM is a branded AI command layer for consulting, delivery, coding, QA, and decision support. It wraps OpenClaw with KAM Solutions theming, a three-agent trio (Kyle · Kade · Knox), and a configuration layout under ~/.kam/ so your installation lives in its own clearly-branded space.
Consulting intelligence. Operational execution. One AI platform.
- Why KAM
- The KAM agents
- Requirements
- Install
- Quick start
- Commands
- State & config location
- Channels — Telegram
- Plugins & MCP
- Customizing the agents
- Troubleshooting
- FAQ
- Security
- Contributing
- License
Unlike consumer AI tools that focus on conversation alone, KAM is positioned as a working layer for professionals — it transforms ideas into deliverables, requirements into systems, decisions into action. It's designed for teams that need:
- Structure complexity — requirements, workflows, policies, governance models
- Accelerate delivery — from idea to implementation faster
- Improve quality — QA discipline, UAT support, validation logic, defect detection
- Automate knowledge work — business analysis, documentation, reporting
- Support better decisions — executive dashboards, benchmarking, decision support
- Operationalize AI in real enterprise environments
Grounded in real enterprise work — digital transformation, business analysis, governance, PMO systems, executive decision support — across healthcare, government, finance, and enterprise systems.
KAM ships with three coordinated agents that mirror how real delivery teams work:
| Agent | Role | Handles |
|---|---|---|
| Kyle 🧠 | Strategic coordination & executive support | BRDs, exec emails, meeting notes, orchestration, research, proposals |
| Kade 🛠️ | Technical implementation | Code, APIs, dashboards, integrations, automation, architecture |
| Knox 🛡️ | Quality assurance & validation | Test cases, UAT, defect detection, requirements review, readiness |
Kyle is the coordinator and delegates to Kade and Knox when a task fits their role. Each agent has its own workspace, identity, and operating instructions — fully customizable via plain markdown files.
- Node.js 20+ (
node --version) - OpenClaw installed globally (
npm i -g openclaw) — KAM uses OpenClaw as its runtime. - A model backend — either Claude Code CLI (free with your Claude subscription) or an Anthropic API key for the embedded backend.
npm i -g openclaw # KAM's runtime dependency
npm i -g kam-cli # the KAM CLI (command installed as `kam`)Note: the npm package is called
kam-clibecause the short namekamwas taken. The command you run remainskam.
kam startFirst run will:
- Copy any existing
~/.openclaw/openclaw.jsoninto~/.kam/openclaw.json - Run the interactive onboarding wizard
- Launch the gateway on
http://127.0.0.1:18789/
Open the dashboard — KAM is live.
# Talk to each agent
kam kyle "Draft an executive update on the Q2 BRD rollout"
kam kade "Add a /health endpoint to the FastAPI service at src/api/app.py"
kam knox "Review BRD v0.3 for completeness and flag any weak requirements"
# Generic form
kam agent <id> <message>
# Gateway
kam gateway status
kam gateway start
kam gateway stop
kam gateway run-bg # background process (useful on Windows if scheduled-task mode misbehaves)
# Telegram
kam telegram add <BOT_TOKEN>
kam telegram list
kam telegram status
kam telegram pair <CODE>
# Config
kam config file
kam config get <path>
kam config set <path> <value>
kam config validate
# Observability
kam doctor
kam doctor --fix
kam where # resolved paths
kam version # kam + openclaw + node + platform
# Escape hatch — any raw openclaw arg, scoped to ~/.kam
kam raw -- plugins list
kam raw -- skills install openclaw-github-assistantKAM stores everything under ~/.kam/:
~/.kam/
├── openclaw.json # main config (treat as secrets)
├── agents/ # per-agent state
├── workspaces/ # per-agent workspace (Kyle / Kade / Knox)
├── extensions/ # installed plugins (e.g., Maton, Brave Search)
├── credentials/ # OAuth/API tokens
└── gateway.log # when using `kam gateway run-bg`
KAM sets OPENCLAW_STATE_DIR=~/.kam and OPENCLAW_CONFIG_PATH=~/.kam/openclaw.json for every openclaw subprocess, so your KAM installation is fully isolated from any plain OpenClaw install on the same machine.
- Create a bot with @BotFather and copy the token.
kam telegram add <TOKEN>- Restart the gateway:
kam gateway stop && kam gateway run-bg - DM your bot — the first message returns a pairing code.
- Approve it:
kam telegram pair <CODE>
From then on, messages to your bot route to Kyle by default. Reply, and you're talking to your KAM agent.
KAM works with any MCP server — Maton (24+ SaaS apps incl. Google Workspace, Notion, Slack, Salesforce), Brave Search, GitHub tooling, and more.
Package MCP servers as a Claude-format plugin (directory with .claude-plugin/plugin.json + .mcp.json), then install:
kam raw -- plugins install /path/to/my-plugin
kam gateway stop && kam gateway run-bgSee docs/plugins.md for the full plugin format, and docs/agents.md for the agent trio's delegation protocol.
Each agent's instructions live in its workspace as plain markdown:
~/.kam/workspace/— Kyle (main)~/.kam/workspaces/kade/— Kade~/.kam/workspaces/knox/— Knox
Key files per workspace:
| File | Purpose |
|---|---|
IDENTITY.md |
Name, vibe, emoji, avatar |
SOUL.md |
Personality, operating style, delegation rules |
USER.md |
What the agent knows about you |
AGENTS.md |
Workspace conventions, delegation playbook |
Edit any of these; the change takes effect on the next agent turn. No restart needed.
kam doctor # diagnose
kam doctor --fix # attempt auto-fixIf the scheduled-task mode fails on Windows (known quirk), fall back to:
kam gateway run-bgOpenClaw's claude-cli backend only bridges MCPs from plugin bundles (not mcp.servers in the config). If your Maton/Brave/etc. tools don't appear:
- Package them as a Claude-format plugin (see docs/plugins.md)
kam raw -- plugins install /path/to/plugin- Restart the gateway
The MCP servers are loading but Claude Code's one-shot -p mode doesn't wait for slow servers. Use direct node /path/to/index.js instead of npx -y @pkg in your MCP config — global installs start instantly.
If API keys were pasted into chat or logs, rotate immediately:
- Maton: https://maton.ai/api-keys → revoke + regenerate
- Brave Search: https://brave.com/search/api/ → revoke + regenerate
- Telegram: DM BotFather
/revoke→ select bot
Paste new keys via kam raw -- configure (interactive), not on the shell.
Is KAM a fork of OpenClaw? No. KAM is a thin npm wrapper around OpenClaw. OpenClaw stays upstream; KAM adds branding, opinionated defaults, and the Kyle/Kade/Knox agent trio.
Does KAM cost anything? The CLI is MIT-licensed and free. You'll need a model backend — either Claude Code (comes with your Claude subscription) or an Anthropic API key (pay-per-token).
Is my data sent anywhere? Only to the model provider you configure (Anthropic, etc.) and any channels/MCPs you explicitly wire up (Telegram, Maton, Brave Search). KAM itself has no telemetry.
Can I run this on a shared/production server? Not recommended. KAM is designed for a single trusted user on a personal machine. The default "yolo" config gives the agent full OS-user permissions. For production, you'd want a harder sandbox policy.
What about Arabic?
Kyle/Kade/Knox are bilingual — their SOUL.md explicitly supports Arabic + English. The CLI itself is English-only but the docs have an Arabic companion.
~/.kam/openclaw.jsontypically contains API keys and OAuth tokens. It'schmod 600on Unix; keep it out of git and backups that leave your machine.- Never commit your
~/.kam/directory. Our.gitignorealready excludes it. - Report vulnerabilities per SECURITY.md — not as public issues.
Pull requests welcome. See CONTRIBUTING.md for scope, commit style, and development setup. Quick start:
git clone https://github.com/Khalil-am/kam-cli
cd kam-cli
npm install && npm link
npm testMIT. See LICENSE. KAM does not redistribute OpenClaw source — OpenClaw is a peer dependency users install themselves under its own license.
KAM by Kam Solutions — AI-powered consulting, execution, and intelligent operations.