Back to skill
Skillv1.0.0

ClawScan security

Test · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

Scanner verdict

BenignFeb 19, 2026, 9:07 AM
Verdict
benign
Confidence
medium
Model
gpt-5-mini
Summary
The skill's requirements and instructions align with its stated purpose (rendering JSON schemas and using OpenAI to generate schemas), but it's instruction-only (no code or install spec) so the agent will rely on external tools and your OpenAI key being present and trustworthy.
Guidance
This skill is instruction-only and expects a materials-cli (Node) tool plus your OpenAI API key. Before installing or using it: 1) Verify the npm package name and source for 'materials-cli' (who publishes it) and prefer installing only packages you trust; installing global npm packages runs third-party code on your system. 2) Understand that using the 'generate' command will send prompts and possibly schema content to OpenAI — don't supply sensitive data. 3) Provide OPENAI_API_KEY only if you trust the package and want the agent to call OpenAI on your behalf. 4) If you prefer, run the CLI manually in a controlled environment rather than giving the agent autonomous ability to invoke it. If you want a stronger assurance, ask the skill author for the exact npm package name, repository URL, or a vetted install spec before installing.

Review Dimensions

Purpose & Capability
noteThe skill claims to be a 'materials-cli' that renders schemas and uses OpenAI for generation; requiring node and OPENAI_API_KEY is consistent. Minor inconsistency: the registry name is 'Test' while the CLI in SKILL.md is 'materials-cli' (labeling mismatch) and there is no bundled materials-cli code or install spec—so the agent will expect the external CLI to exist on the host or be installed by the user.
Instruction Scope
okSKILL.md only instructs rendering, generating (via OpenAI), and validating JSON schema files. It references schema file paths and optional OpenAI settings; it does not ask the agent to read unrelated files, harvest system secrets, or send data to unexpected endpoints beyond OpenAI.
Install Mechanism
noteThere is no install spec or bundled code; SKILL.md suggests installing 'materials-cli' via npm (-g). That means the agent/user will pull a package from the public npm ecosystem if they follow the README instructions. This is a legitimate, moderate-risk choice (npm packages are common) but the skill itself does not provide a vetted install source or package integrity information.
Credentials
okOnly OPENAI_API_KEY is required (declared as primaryEnv), which matches the documented 'generate' command that uses OpenAI. The SKILL.md also mentions optional OPENAI_MODEL and OPENAI_BASE_URL used if flags are not provided; those are optional and not required environment variables.
Persistence & Privilege
okThe skill does not request permanent presence (always:false) and does not declare modifications to other skills or system settings. Autonomous invocation is allowed (platform default) but is not combined with other concerning privileges.