AI / OpenClaw Interview Questions
1. What is OpenClaw?
OpenClaw is a free, open-source personal AI agent that runs on your own machine and connects large language models directly to your files, shell, browser, and messaging apps like WhatsApp, Telegram, Slack, and Discord. Unlike a typical chatbot that only answers questions in a request-response loo...
2. Who created OpenClaw?
OpenClaw was created by Peter Steinberger, an Austrian software developer previously known as the founder of PSPDFKit. The project began as a personal weekend tool and grew rapidly after its public release Reporting has described Steinberger moving toward OpenAI, with OpenClaw itself transitionin...
3. What is OpenClaw's license?
OpenClaw is released under the MIT License, a permissive open-source license. The software itself is completely free to use, with no subscription fee Costs instead come from the AI model API calls it makes, since users bring their own API keys for providers like Anthropic or OpenAI Being open-sou...
4. What are the earlier names OpenClaw was known by?
OpenClaw wasn't launched under its current name, it went through a few rebrands as the project grew. It started as a personal tool called Clawd, named after Anthropic's Claude It was then released publicly under the name Clawdbot in November 2025 It was also referred to as Moltbot at one point be...
5. What is the Gateway in OpenClaw's architecture?
The Gateway is the central component of OpenClaw's architecture, acting as a self-hosted control plane that manages communication, session continuity, access control, and agent execution. Routes requests, described internally as node.invoke frames, from the Agent Runtime to the appropriate local ...
6. What is a Skill in OpenClaw?
A Skill is OpenClaw's plugin mechanism, an extension that gives the agent a new capability, such as interacting with a specific external service, without modifying OpenClaw's own source code. Defined through a SKILL.md instruction file plus supporting scripts Can be installed from the public Claw...
7. What is ClawHub?
ClawHub is OpenClaw's public registry for discovering and installing community-built Skills. Lets developers publish and share skills that extend what an OpenClaw agent can do Some installations ship with over a hundred prebuilt skills, and developers can add their own on top Security researchers...
8. What is Canvas in OpenClaw?
Canvas is an agent-driven visual workspace within OpenClaw that runs as its own separate server process, rather than being folded into the main Gateway. Defaults to running on port 18793 Kept isolated from the Gateway specifically so that if Canvas crashes, the Gateway keeps operating normally Re...
9. What messaging channels does OpenClaw support?
OpenClaw is built to connect to the communication channels people already use, rather than requiring a dedicated app. Supports WhatsApp, Telegram, Slack, Discord, Signal, iMessage, and WebChat directly Some sources describe support extending to 30 or more platforms in total Also supports local an...
10. What AI model providers does OpenClaw support?
OpenClaw is model-agnostic rather than being locked to a single AI provider. Supports Anthropic's Claude and OpenAI's GPT models through their respective APIs Supports local models run through Ollama, letting inference happen entirely on the user's own hardware Uses a model-selection layer that r...
11. What are the minimum hardware requirements to run OpenClaw?
OpenClaw's hardware needs scale depending on whether it's calling a cloud model or running inference locally. Setup Recommended resources Absolute minimum 2 GB RAM and 2 CPU cores Using cloud-based models 4 to 8 GB RAM recommended Running local LLMs 16 to 24 GB RAM recommended The gap between the...
12. What is Sandbox Mode in OpenClaw?
Sandbox Mode controls whether OpenClaw isolates its execution environment, such as through Docker containers, or runs with direct access to the host system. Disabled by default, meaning out of the box, an OpenClaw agent runs with the same permissions as the user account it's launched under Can be...
13. Define the Agent Runtime component in OpenClaw?
The Agent Runtime is the component that actually executes an agent's reasoning: the language model reasoning loop, tool dispatch, and, when enabled, Docker sandbox management. Resolves authentication profiles and selects which model to use for a given run Submits conversation turns to the LLM pro...
14. What is a SKILL.md file in OpenClaw?
A SKILL.md file is the instruction manifest that defines a Skill in OpenClaw, describing what the skill does and what capabilities or permissions it declares needing. Loaded into the agent's context so the model knows the skill exists and how to use it The declared permissions in this manifest pr...
15. What is the default port for OpenClaw's web gateway?
OpenClaw's web gateway listens on port 18789 by default. If this port is bound to all network interfaces rather than a private or loopback address, it becomes reachable from the internet Security researchers have identified tens of thousands of internet-exposed OpenClaw instances using this kind ...
16. What is the default port for OpenClaw's Canvas process?
Canvas, OpenClaw's agent-driven visual workspace, defaults to running on port 18793, separate from the Gateway's own port. Running as its own server process rather than sharing the Gateway's port reinforces the isolation between the two components Means a problem in Canvas doesn't necessarily exp...
17. Describe how OpenClaw differs from a traditional chatbot?
A traditional chatbot like a standard ChatGPT-style assistant operates in a stateless request-response loop, you send a prompt, it replies, and the interaction is over. OpenClaw instead runs continuously, holding context and session state across an ongoing task rather than just a single exchange ...
18. What is the purpose of the agents.list configuration?
The agents.list configuration lets an OpenClaw deployment define multiple distinct agents, each with its own workspace, tool policy, and model, rather than running one single, all-purpose agent. A sales assistant agent might have CRM tools enabled while a DevOps-focused agent has server scripts e...
19. How do you install a skill from ClawHub?
Skills are typically installed and managed through the clawhub CLI, which pulls a skill from the registry into an agent's workspace. In team settings, skills can be centrally curated, with approved skills loaded automatically from a shared workspace directory Security guidance strongly recommends...
20. What is Trusted-Proxy authentication in OpenClaw?
Trusted-Proxy authentication is how OpenClaw supports enterprise single sign-on without shipping native identity provider connectors of its own. A reverse proxy in front of OpenClaw handles the actual OAuth, OIDC, or SAML authentication flow That proxy forwards authenticated identity headers to O...
21. List the four architectural layers of OpenClaw's hub-and-spoke design?
Security analysis of OpenClaw describes its architecture as a hub-and-spoke design made up of four layers. Layer Role Channel adapters Connect OpenClaw to messaging platforms like WhatsApp, Telegram, Slack, and Discord Gateway control plane Routes requests, manages sessions, and brokers agent exe...
22. What does OpenClaw cost to run?
OpenClaw the software is entirely free, since it's open-source and MIT licensed with no subscription fee attached. Actual costs come from the AI model API calls it makes on your behalf, since you bring your own API key Light usage tends to run roughly ten to thirty dollars a month, typical usage ...
23. What is session serialization in OpenClaw?
By default, OpenClaw processes messages within a single session one at a time, in order, rather than handling multiple messages from the same conversation simultaneously. Prevents a session's state from being corrupted by two messages being processed concurrently and racing against each other Doe...
24. Define the Tool Layer in OpenClaw's architecture?
The Tool Layer is the architectural component responsible for actually carrying out the actions an agent decides to take, file access, shell commands, browser automation, and calls to external services. Receives dispatched tool calls from the Agent Runtime once the model decides an action is need...
25. What is Skill Discovery vs Skill Injection in OpenClaw?
OpenClaw distinguishes between simply knowing a skill exists and actually loading its full instructions into the model's context for a given turn. Skill discovery : OpenClaw can identify what skills are available at runtime Skill injection : only the skill or skills relevant to the current task a...
26. What is the difference between Sandbox Mode off and Docker-based sandboxing in OpenClaw?
Sandbox Mode off Docker-based sandboxing Agent runs with the same permissions as the host user account Agent runs inside an isolated container with restricted OS-level access Lower friction, no container overhead Adds isolation at the cost of some setup complexity A compromise affects the entire ...
27. Why does OpenClaw selectively inject only relevant skills instead of loading all skills into context?
If every installed skill's full instructions were injected into every single prompt, the context window would fill up quickly with mostly irrelevant information. A bloated context makes it harder for the model to focus on what's actually relevant to the current task Unused injected content still ...
28. How does OpenClaw achieve data sovereignty when running local models via Ollama?
Data sovereignty means keeping data, and the processing of that data, entirely within infrastructure an organization controls, rather than sending it to an external provider. OpenClaw's model-selection layer can be configured to route requests to a local model running through Ollama instead of a ...
29. What is the difference between OpenClaw's multi-agent support and running a single agent instance?
Single agent instance Multi-agent setup via agents.list One shared workspace, tool policy, and model for everything Each agent has its own workspace, tool policy, and model Simpler to configure and reason about More setup, but enables role-based permission separation A compromise or mistake affec...
30. Why is a skill in OpenClaw treated as a trusted instruction source by the Agent Runtime?
Once a skill is loaded, its SKILL.md instructions become part of what the Agent Runtime treats as legitimate guidance for how the agent should behave, the same level of trust given to the agent's own core configuration. This is what lets a skill meaningfully extend the agent's behavior, since the...
31. How does OpenClaw's per-agent tool policy limit the blast radius of a compromised agent?
Because each agent under a multi-agent configuration only has the specific tools declared in its own policy, a compromise of one agent doesn't automatically grant access to every capability across the whole deployment. An agent with CRM tools enabled has no path to server-script execution if that...
32. What is the difference between OpenClaw's Gateway and its Canvas process?
Gateway Canvas Central control plane handling routing, sessions, and access control Agent-driven visual workspace for rendering agent-generated content Runs on port 18789 by default Runs on port 18793 by default, as a separate process Failure affects the whole platform's core routing Failure is i...
33. Why do security researchers recommend OAuth over long-lived API keys for OpenClaw skill integrations?
A long-lived API key is a static credential, if it leaks, it typically remains valid and usable until someone manually notices and revokes it. OAuth instead issues scoped, time-limited authorization tokens rather than a permanent static secret A leaked OAuth token naturally expires and is limited...
34. What is the difference between OpenClaw serializing messages per session and running agents in parallel across sessions?
These two behaviors operate at different scopes and aren't in tension with each other. Per-session serialization : messages within the same conversation are processed one at a time, in order, keeping that specific session's state consistent Cross-session parallelism : separate, unrelated sessions...
35. Why was OpenClaw banned from some corporate work devices?
Some organizations, including reports naming Meta, have restricted OpenClaw on work devices due to concerns about its unpredictability and potential for privacy breaches in otherwise secure environments. OpenClaw's defining feature, direct system access combined with autonomous multi-step executi...
36. How does exposing OpenClaw's gateway port to 0.0.0.0 create a security risk?
Binding the Gateway's web interface to 0.0.0.0 means it listens on every network interface on the machine, including ones reachable from the public internet, rather than staying restricted to a private or loopback address. Anyone who discovers the server's IP address can reach the OpenClaw interf...
37. What is the difference between a skill's self-declared permissions and independently verified sandbox enforcement?
Self-declared permissions Independently verified enforcement Listed in the skill's SKILL.md manifest by its author Enforced at runtime regardless of what the manifest claims Provides capability transparency, what the skill says it needs Provides an actual guarantee about what the skill can do Tru...
38. Why does an audit log stored in a mutable logs directory undermine forensic attribution?
An audit log is only useful for investigating an incident if it can be trusted to accurately reflect what actually happened. If the log lives in a directory the agent itself can write to, a compromised agent, or a malicious skill running with the agent's permissions, could overwrite or delete rel...
39. How does OpenClaw route a request through its model-selection layer?
The model-selection layer decides which LLM provider actually handles a given request, rather than the choice being hardcoded to a single provider. Considers the user's configuration, such as a preferred provider for a specific agent or task type Checks provider availability, since a failover att...
40. What is the difference between running OpenClaw directly on a host system versus inside a sandbox?
Direct on host Inside a sandbox Agent runs with the full permissions of the launching user account Agent's access is constrained to whatever the sandbox boundary allows A compromise can reach files, SSH keys, and anything else that user can access A compromise is contained within the sandbox in m...
41. Explain the lifecycle of a task from a chat message to a completed OpenClaw agent response?
flowchart LR A[Message Sent via WhatsApp/Telegram/etc] --> B[Channel Adapter] B --> C[Gateway Control Plane] C --> D[Agent Runtime: reasoning loop] D --> E[Tool Dispatch] E --> F[Tool Execution: sandboxed if configured] F --> D D --> G[Response Delivery via Channel Adapter] G --> H[Session State ...
42. Explain the internal working of the Gateway's node.invoke frame routing?
sequenceDiagram participant CA as Channel Adapter participant GW as Gateway participant AR as Agent Runtime participant LE as Local Execution CA->>GW: incoming message GW->>AR: node.invoke frame AR->>LE: tool dispatch LE-->>AR: result AR-->>GW: response GW-->>CA: deliver Internally, OpenClaw's Ga...
43. How can you optimize OpenClaw's context window usage across a session with many available skills?
Rely on selective skill injection rather than manually forcing every skill into context, since OpenClaw already avoids injecting irrelevant skills by default Prune unused or rarely-needed skills from an agent's workspace rather than leaving every ever-installed skill discoverable Use multi-agent ...
44. How do you troubleshoot an OpenClaw agent that appears to hang or serialize incorrectly across concurrent sessions?
Confirm whether the behavior is happening within a single session or across multiple sessions, since per-session serialization is expected behavior, not a bug, for messages within the same conversation Check whether the Agent Runtime's failover attempt loop is repeatedly retrying against an unava...
45. Which is better and why: running OpenClaw sandboxed inside Docker vs running it directly on a trusted personal device?
Sandboxed inside Docker Directly on a trusted personal device Contains a compromise within the sandbox boundary in most cases A compromise has the same reach as the launching user account Added setup and some operational overhead Lowest friction, fastest to get running Well suited to shared infra...
46. Explain the internal working of OpenClaw's Docker sandbox management within the Agent Runtime?
flowchart LR A[Tool Call Requested] --> B[Agent Runtime checks sandbox.mode] B -->|Enabled| C[Spin up Docker Container] C --> D[Execute Tool Call inside Container] D --> E[Capture Result] E --> F[Tear Down or Reuse Container] F --> G[Return Result to Reasoning Loop] B -->|Disabled| H[Execute Dire...
47. Explain, at a high level, the CVE-2026-25253 vulnerability and how OpenClaw addressed it?
CVE-2026-25253 was a high-severity vulnerability, rated 8.8 on the CVSS scale, that allowed remote code execution triggered through a malicious webpage, requiring only a single click of interaction from the user. The vulnerability involved how the agent's browser-automation tooling processed cont...
48. Explain the lifecycle of a compromised skill's permission risk, from installation to potential impact?
flowchart LR A[Skill Published to ClawHub] --> B[Skill Installed via clawhub CLI] B --> C[SKILL.md Loaded into Agent Context as Trusted Instructions] C --> D[Skill Executes with Agent's Full Granted Permissions] D --> E[Self-Declared Permissions Not Independently Verified] E --> F[Potential Later...
