The Agent Skills Directory

JavaScript Execution Capability: The primary function of this skill is to generate and execute JavaScript code within Figma design files. This represents a high-capability tool used for complex design automation tasks. The skill provides extensive educational documentation to ensure the generated code is syntactically and logically correct according to Figma's API specifications.
Network Isolation: A significant security feature documented in the skill is the lack of network access for the use_figma tool. This prevents the possibility of data exfiltration (e.g., design assets or file structure) to external servers, even if a script were compromised or malicious.
Indirect Prompt Injection Surface: The skill guides the agent to read existing file content (text, metadata, and variables). As with any skill that ingests untrusted data from a file, there is a theoretical risk of indirect prompt injection if a Figma file contains text designed to manipulate the agent's behavior. However, this is an inherent risk of the task itself rather than a flaw in the skill.
Trusted Vendor Context: The skill is authored by a trusted vendor (openai) and facilitates interaction with established, well-known services (Figma). All referenced domains (figma.com, figma.io) and API patterns are legitimate and standard for the Figma ecosystem.
Safety Guardrails: The documentation includes numerous 'Gotchas' and rules aimed at preventing common API failures, such as proper font loading, coordinate handling, and color range validation (0-1), which contributes to the overall stability and safety of the tool's usage.

figma-use