Back to skill
Skillv0.1.0
ClawScan security
Brand Guidelines · ClawHub's context-aware review of the artifact, metadata, and declared behavior.
Scanner verdict
BenignMar 10, 2026, 2:31 AM
- Verdict
- benign
- Confidence
- high
- Model
- gpt-5-mini
- Summary
- This is an instruction-only branding/style guide that matches its name and asks for no credentials, installs, or privileged access.
- Guidance
- This skill is purely an instruction/STYLE doc and appears safe to install: it asks for no credentials and writes nothing. Notes before using: (1) the guide references using python-pptx and system fonts—if you want automated styling, install python-pptx and the Poppins/Lora fonts from trusted sources; (2) the LICENSE is Apache-2.0 but trademark use of Anthropic branding may require permission—ensure you are authorized to apply Anthropic's brand to artifacts; (3) if an implementation is added later (scripts that run automatically), re-evaluate for network calls, file access, or required credentials before allowing execution.
Review Dimensions
- Purpose & Capability
- okThe name and description (applying Anthropic brand colors/typography) match the SKILL.md content. The skill is documentation-only and does not request unrelated capabilities or credentials.
- Instruction Scope
- okSKILL.md contains only guidance about colors, fonts, and recommended application (mentions python-pptx for RGBColor), with no commands, file reads, network endpoints, or instructions to collect or transmit data. It stays within the stated purpose.
- Install Mechanism
- noteThere is no install spec (lowest risk). The doc references python-pptx but does not declare or install it; this is a minor documentation note rather than a security problem—if you expect automated application, you or the execution environment will need to install that library from a trusted source.
- Credentials
- okThe skill requests no environment variables, credentials, or config paths. Nothing disproportionate is asked for or required.
- Persistence & Privilege
- okThe skill is not always-enabled, does not request persistent presence, and does not modify other skills or system settings.
