The Agent Skills Directory

Indirect Prompt Injection Surface: The skill involves reading and processing content from external Figma files, which is an expected operational pattern for design automation.
Ingestion points: Data enters the agent context through read-only scripts and visual inspection of Figma Slides files, as seen in SKILL.md and references/slide-gotchas.md.
Boundary markers: The instructions do not explicitly mandate the use of delimiters for processing retrieved text content, though this is a standard design consideration for such tools.
Capability inventory: The agent has the capability to create, modify, and delete nodes within the Figma environment using the use_figma tool, as described in references/slide-lifecycle.md and references/slide-content.md.
Sanitization: The skill relies on standard platform validation and model-level guardrails for processing extracted content.
Managed Asset Uploads: The documentation describes a specific workflow for image management using authorized tool endpoints. This process is contained within the Figma ecosystem and uses temporary, single-use URLs for asset commitment, which is consistent with standard platform security practices.
Technical Implementation Standards: The provided code examples and references, such as those in references/slide-gotchas.md, focus on addressing technical constraints of the Slides API (e.g., coordinate management) and do not involve any unauthorized system access or command execution.

figma-use-slides