← Back to Skills Marketplace
580
Downloads
2
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install worldbook
Description
AI's Knowledge Base CLI - Query and manage world knowledge for AI agents. Use when users want to search knowledge, add knowledge sources, or interact with th...
Usage Guidance
This skill is plausible for its stated purpose but has two practical risks you should consider before installing/using it: (1) The SKILL.md assumes and recommends installing a third‑party 'worldbook' CLI from PyPI/npm/GitHub, yet the skill metadata doesn't declare that dependency — verify the package and repository (author, release history, source code) before running any install. (2) The skill instructs agents to fetch external 'worldbook' documents and inject them into agent context; treat such external instructions as untrusted input. If you deploy this skill, restrict the agent from auto-installing packages or auto‑executing fetched instructions, run installs in a sandbox, review worldbook content before injection, and prefer pinned package versions or vetted sources. If you want to proceed, ask the publisher for the official package name, repository verification (checksums/signatures), and a declared required-binary/install spec in the metadata to remove the coherence gap.
Capability Analysis
Type: OpenClaw Skill
Name: worldbook
Version: 0.1.0
The `SKILL.md` describes a core functionality where the AI agent is instructed to fetch 'instructions for AI' from an external source (via `worldbook get <name>`) and then explicitly 'inject it into your context'. This design pattern creates a severe prompt injection vulnerability, as it allows arbitrary, externally controlled instructions to be executed by the AI agent. While the stated purpose is benign ('learn how to use the service'), this mechanism provides a direct channel for an attacker to compromise the agent's behavior if the external 'worldbook' content source is compromised, leading to potential unauthorized actions or data exfiltration.
Capability Assessment
Purpose & Capability
The skill is marketed as a CLI-first knowledge-base (worldbook) but the registry metadata declares no required binaries or install steps. SKILL.md assumes a worldbook CLI exists (and shows pip/npm/git install commands). Not declaring the CLI binary or an install spec in the skill metadata is an incoherence: either the skill should include/declare the CLI or it should be explicit that the skill is instruction-only and won't function without a separately installed package.
Instruction Scope
Runtime instructions tell agents to run 'worldbook get' and to 'inject' returned text into context. That means the agent will fetch and absorb arbitrary external instructions/text; the SKILL.md also recommends installing packages from PyPI/npm or cloning a GitHub repo. These steps can cause an agent to fetch and execute untrusted code or to incorporate unvetted instructions that could lead to undesired actions. The instructions do not include safety constraints or validation steps for external worldbook content.
Install Mechanism
There is no install spec in the skill metadata, but SKILL.md recommends installing 'worldbook' via pip, npm, or a git clone (https://github.com/femto/worldbook-cli). Suggesting package installs from public registries is common, but because the skill metadata omitted this dependency, it's not clear whether the platform or the user is expected to vet or sandbox those installs. Installing third-party CLI packages can execute arbitrary code on the host — the skill gives direct install commands without provenance/verification guidance.
Credentials
The skill declares no required environment variables, credentials, or config paths. There are no mismatched or excessive secret requests in the metadata or SKILL.md.
Persistence & Privilege
The skill does not request always:true or other elevated persistence. It uses default autonomous-invocation behavior (normal). The skill does not declare changes to other skills or system-wide config.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install worldbook - After installation, invoke the skill by name or use
/worldbook - Provide required inputs per the skill's parameter spec and get structured output
Version History
v0.1.0
worldbook 0.1.0 initial release
- Introduces a CLI tool for querying and managing a shared AI knowledge base.
- Allows agents to search for, fetch, and add worldbook entries via simple commands.
- Provides installation options for Python and Node.js.
- Focuses on a CLI-first approach—enabling structured, machine-readable information for AI agents.
- Includes examples and usage philosophy in the initial documentation.
Metadata
Frequently Asked Questions
What is Worldbook?
AI's Knowledge Base CLI - Query and manage world knowledge for AI agents. Use when users want to search knowledge, add knowledge sources, or interact with th... It is an AI Agent Skill for Claude Code / OpenClaw, with 580 downloads so far.
How do I install Worldbook?
Run "/install worldbook" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Worldbook free?
Yes, Worldbook is completely free (open-source). You can download, install and use it at no cost.
Which platforms does Worldbook support?
Worldbook is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Worldbook?
It is built and maintained by femto (@femto); the current version is v0.1.0.
More Skills