seed-migrate-component-docs-from-figma

Warn

Audited by Gen Agent Trust Hub on Sep 1, 2026

Risk Level: MEDIUMDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [DYNAMIC_EXECUTION]: The skill uses a user-supplied file-name to construct the target file path (docs/content/components/{file-name}.mdx). There is no validation or sanitization of this input, which could allow an attacker to perform directory traversal (e.g., using ../../) to write files to unauthorized locations on the filesystem.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted text content from external Figma layers via the get_node_info tool and writes it to documentation files. The lack of sanitization or boundary markers for this content creates an attack surface where malicious instructions or payloads within Figma could influence the agent or be persisted into the codebase.
  • Ingestion points: Extracted text content from Figma node URLs in SKILL.md.
  • Boundary markers: No explicit delimiters or instructions are used to separate external Figma content from agent instructions.
  • Capability inventory: The skill possesses file system write capabilities and access to Figma data retrieval tools.
  • Sanitization: No filtering, escaping, or validation of the ingested Figma content is performed before it is written to the MDX files.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Sep 1, 2026, 05:48 AM
Security Audit — agent-trust-hub — seed-migrate-component-docs-from-figma