Skillv2.0.4

ClawScan security

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

Scanner verdict

BenignMar 18, 2026, 8:29 AM

Verdict: Benign
Confidence: high
Model: gpt-5-mini
Summary: The skill is an instruction-only design critique tool whose declared requirements, instructions, and files are coherent with its stated purpose and do not request unrelated credentials or installs.
Guidance: This skill appears coherent and low-risk because it's instruction-only and requests no credentials or installs. Before installing/using, consider: (1) Privacy — images can contain sensitive data (user faces, screenshots with API keys, personal info). Avoid uploading anything private or confidential unless you trust where image data is processed. (2) Figma / URLs — the skill mentions Figma frames and website URLs but does not request Figma tokens or provide a fetch mechanism; you must supply images/screenshots or allow the agent to fetch URLs (which may trigger browsing/network actions by the agent). (3) Processing location — confirm whether your platform sends uploaded images to third-party models/endpoints for vision analysis; the skill itself declares no external endpoints. (4) Autonomous agents — while this skill is not requesting extra privileges, if you allow autonomous agent execution, monitor any agent actions that fetch external URLs or request credentials. If any of the above is a concern, avoid uploading sensitive images or require explicit confirmation before the agent fetches content.

Review Dimensions

Purpose & Capability: okName/description (design critique) align with the content: the SKILL.md describes image-based critique using design frameworks and the bundled reference docs support that. The skill declares no binaries, env vars, or installs — which is proportionate because it operates as an instruction-only critique template that relies on the agent's existing image/vision capabilities or user-supplied images/URLs.
Instruction Scope: okInstructions stay within the stated domain: they ask the agent to evaluate user-provided visuals across 11 frameworks and produce prioritized findings with fixes. The SKILL.md does not instruct the agent to read unrelated files, environment variables, or system paths, nor to transmit data to unexpected endpoints. Note: it accepts website URLs and references Figma frames, but provides no built-in fetching mechanism or credential requests — it therefore relies on the agent or user to supply those visuals, which is consistent but worth noting.
Install Mechanism: okNo install spec or code files are provided (instruction-only). That minimizes the risk of arbitrary code being written to disk or external downloads. This is appropriate for a critique/template skill.
Credentials: okThe skill requires no environment variables, secrets, or external credentials. The lack of requested credentials is proportionate to a design critique capability and avoids unnecessary access to cloud or design platform accounts. (If users expect the skill to fetch Figma frames automatically, they should note the skill does not request Figma tokens.)
Persistence & Privilege: okThe skill does not request always:true or any system-wide persistence; it is user-invocable and uses the platform defaults for autonomous invocation. There is no evidence it modifies other skills or system settings.