The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill facilitates the installation of the @paragraph-com/sdk package via npm to enable programmatic access to the Paragraph platform.
[COMMAND_EXECUTION]: The skill utilizes curl for REST API requests and npm for SDK management, which are legitimate operations for an API integration tool.
[PROMPT_INJECTION]: The skill retrieves and processes external content (e.g., markdown from blog posts, feeds, and search results) from Paragraph.com, creating an attack surface for indirect prompt injection.
Ingestion points: Untrusted data enters the agent context through api.posts.list, api.posts.get, api.feed.get, and api.search.posts in SKILL.md.
Boundary markers: The skill instructions do not specify the use of delimiters or instructions to ignore embedded commands within the fetched content.
Capability inventory: The skill's environment includes Write permissions and Bash execution, which are potentially exploitable if the agent obeys instructions embedded in the external markdown.
Sanitization: There is no evidence of sanitization, escaping, or validation of the external content before it is processed or written to the local filesystem.

paragraph-api