The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The installation instructions in SKILL.md use a dangerous pattern that pipes a remote script directly from v3.heroui.com into the bash shell (curl -fsSL https://v3.heroui.com/install | bash). This executes arbitrary code from a remote server without prior inspection.
[EXTERNAL_DOWNLOADS]: Several migration scripts (e.g., get_component_migration_guides.mjs, get_migration_guide.mjs) fetch MDX-formatted documentation from a staging environment hosted on Vercel (heroui-git-docs-migration-heroui.vercel.app). While Vercel is a trusted provider, fetching content from non-production staging URLs introduces risks regarding the stability and integrity of the source.
[PROMPT_INJECTION]: The skill exhibits a significant surface for indirect prompt injection (Category 8) because it retrieves documentation from external URLs and displays it as authoritative guides for the agent to follow during code migration.
Ingestion points: MDX files fetched by the Node.js scripts in the scripts/ directory.
Boundary markers: The skill does not provide any clear delimiters or instructions for the agent to ignore embedded commands within the fetched documentation.
Capability inventory: The agent is expected to use these guides to perform code modifications, change component APIs, and update project dependencies.
Sanitization: The scripts do not perform any sanitization or validation of the remote content before presenting it to the agent.

heroui-migration