claude2figma

Warn

Audited by Gen Agent Trust Hub on Jun 13, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill initiates the download and installation of a package from 'github.com/senlindesign/claude2figma', which is not a recognized trusted vendor or well-known service organization.
  • [REMOTE_CODE_EXECUTION]: It instructs the user to run 'npx skills add senlindesign/claude2figma', a command that downloads and executes external code from the internet to modify the agent's environment.
  • [COMMAND_EXECUTION]: The skill provides shell commands to check for existing installations in the home directory ('~/.design-agent-skills/skills/') and suggests using the shell escape character ('!') in Claude Code to execute these commands.
  • [PROMPT_INJECTION]: The skill processes data from Figma canvas operations to verify compliance, which creates a potential surface for indirect prompt injection.
  • Ingestion points: Figma design data, including component instances, color tokens, typography, and spacing values.
  • Boundary markers: Absent; no delimiters or instructions to ignore embedded content are provided for the ingested design data.
  • Capability inventory: Access to shell command execution (via npx), file system access in the home directory, and Figma canvas write capabilities.
  • Sanitization: No sanitization or validation of external design data is mentioned.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jun 13, 2026, 10:44 AM
Security Audit — agent-trust-hub — claude2figma