The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to fetch and process external data such as design briefs and asset manifests to automate asset generation.
Ingestion points: The skill fetches logos, images, and manifest files from URLs or paths provided in the user's brief as described in SKILL.md.
Boundary markers: The instructions lack explicit delimitation or warnings to ignore instructions that might be embedded within the processed assets or manifests.
Capability inventory: The skill utilizes file system writes for asset export and utilizes Python for rendering.
Sanitization: No sanitization or validation logic is specified for the input data before it influences the asset generation workflow.
[DYNAMIC_EXECUTION]: The instructions direct the agent to "Build a deterministic renderer" using Python and the Pillow library in SKILL.md.
This involves the dynamic generation and execution of code based on parameters found in the asset brief. Generating scripts at runtime based on external input creates a potential path for logic injection into the rendering environment.
[EXTERNAL_DOWNLOADS]: The skill references and encourages loading resources from an external source, specifically the TypeUI/awesome-design-skills repository under the bergside organization in references/typeui-design-style-skills.md.
The agent is instructed to load these external systems to define typography, color tokens, and layout rules, introducing a dependency on third-party managed content not originating from the skill author.
[COMMAND_EXECUTION]: The reference material in references/typeui-design-style-skills.md suggests using the open command on macOS to show results to the user.
This involves executing shell commands to interact with the host operating system's file handler.

static-social-campaign-assets