A comprehensive security framework adding eight governance layers to protect and monitor your AI agent operations.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install samma-suit
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install samma-suit using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Samma Suit Security Governance?
Samma Suit is an open-source security framework specifically designed to wrap AI agents in a robust governance layer. By intercepting plugin hooks, it ensures that every action taken by the agent is vetted against strict security policies. This skill is essential for developers using Openclaw Skills who need to move beyond experimental setups into production-grade environments where safety, cost management, and auditability are non-negotiable.
By providing a structured way to manage agent permissions and resource consumption, Samma Suit mitigates the risks associated with autonomous tool execution. It functions as a security gatekeeper, ensuring that every interaction within the ecosystem of Openclaw Skills adheres to predefined ethical and operational boundaries.
Samma Suit Security Governance Use Cases
- Implementing a kill switch to immediately terminate agent activity during detected anomalies.
- Enforcing monthly spending caps to prevent unexpected API costs across multiple agents.
- Vetting external skills via AST scanning to block unapproved or malicious code execution.
- Maintaining a cryptographic audit trail for every message and tool call made by the agent for regulatory compliance.
How Samma Suit Security Governance Works
- The framework intercepts OpenClaw plugin hooks during the agent's lifecycle to evaluate incoming requests.
- It validates the current request against the DHARMA permission set to ensure the tool call is authorized.
- The KARMA layer checks the remaining monthly budget before allowing API-heavy operations to proceed.
- The SANGHA layer performs AST scanning on any external skills being invoked to verify they are on the allowlist.
- METTA signs outbound messages with Ed25519 keys to provide cryptographic proof of the agent's identity.
- SILA records the entire transaction history to a persistent audit log for post-session analysis.
Samma Suit Security Governance Setup
To install Samma Suit, you must have your API key ready. Use the following commands to integrate the security suite into your environment and enable it for your Openclaw Skills:
# Install the Samma Suit skill
openclaw install samma-suit
# Configure your required API key
export SAMMA_API_KEY="your_key_here"
Samma Suit Security Governance Data Schema & Taxonomy
Samma Suit organizes data around governance layers to ensure transparency and accountability. Below is the metadata taxonomy used to manage the security state:
| Layer | Data Managed | Primary Purpose |
|---|---|---|
| KARMA | Budget Caps | Financial Governance and usage limits |
| SILA | Event Logs | Audit, traceability, and session history |
| METTA | Ed25519 Keys | Identity verification and message signing |
| DHARMA | Permission Sets | Tool-level access control and authorization |
| BODHI | Resource Limits | Process isolation and token usage timeouts |
Samma Suit Security Governance Advanced Features
- NIRVANA global kill switch for instant termination of all agent activity.
- BODHI process isolation providing hard timeouts and resource limits per agent instance.
- AST-based skill vetting to prevent dynamic code injection from unverified sources.
- Cryptographic non-repudiation for all outbound messages via Ed25519 signing.
SKILL.md
Loading
Related Openclaw Skills
OneMind Skill
A platform for collective alignment where agents and humans collaborate to build consensus through proposition submission and grid-based rating.
OneMind Skill
A specialized skill for AI agents to participate in collective alignment and consensus-building through proposition submission and grid rating.
Komodo Skill
A comprehensive interface for managing Komodo Core resources, enabling automated deployments, monitoring, and infrastructure control.
TickTick Integration
A comprehensive toolkit for AI agents to manage TickTick tasks and projects with automated OAuth2 token handling and typed API access.
TTS via ComfyUI (Qwen3)
An automated workflow to generate high-quality text-to-speech audio using ComfyUI and the Qwen3 model.
Set Reminder
Set Reminder allows AI agents to schedule one-time or recurring notifications delivered through iMessage, Discord, and other channels using the Openclaw Skills cron system.
Featured*
