The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill provides installation instructions for the 'ntn' CLI that involve piping a remote script from an external domain directly into the bash shell.
Evidence: curl -fsSL https://ntn.dev | bash in SKILL.md.
Risk: Piped execution of remote scripts from unverified sources bypasses inspection and can lead to full system compromise.
[INDIRECT_PROMPT_INJECTION]: The skill retrieves and processes external content from the Notion API, creating a vector for instructions embedded in pages or databases to influence agent behavior.
Ingestion points: Methods like ntn api v1/pages/{page_id}/markdown and curl -s "https://api.notion.com/v1/pages/{page_id}/markdown" pull external text into the agent context.
Boundary markers: Absent. There are no instructions defining how to delimit external data from agent instructions.
Capability inventory: The skill has the ability to execute shell commands (ntn, curl) and perform network writes to the Notion API.
Sanitization: No evidence of content filtering or escaping was found for data retrieved from external API endpoints.
[COMMAND_EXECUTION]: The skill relies on the execution of multiple CLI tools to perform its primary functions.
Evidence: Frequent use of ntn and curl for API interactions and file management as documented in SKILL.md.

notion