The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill downloads design content from the Open Design repository hosted on GitHub (https://github.com/nexu-io/open-design) using standard git clone and git pull commands.
[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands, specifically git for repository management and node to run local helper scripts (e.g., scripts/list-design-systems.mjs) for parsing metadata.
[INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection by ingesting and acting upon instructions from external markdown files.
Ingestion points: The agent is instructed to read DESIGN.md, SKILL.md, and craft/*.md files from the local clone of the Open Design repository (typically located at ~/.open-design-skill/repo).
Boundary markers: There are no explicit delimiters or instructions to ignore embedded commands within the fetched markdown files before they are incorporated into the agent's active workflow.
Capability inventory: The agent has permissions to execute shell commands (git, node), perform file system operations (read/write), and interact with the user via specific questions.
Sanitization: While the helper scripts use regular expressions to extract structured metadata for display, the final execution phase involves reading the full content of remote markdown files and following their instructions without additional validation or sanitization.

open-design