The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill instructions and documentation (in README.md and manifest.json) recommend installing the Notion CLI dependency using a shell pipe: curl -fsSL https://ntn.dev | bash. This pattern downloads and executes a script directly in the user's environment, which is the official installation method for the Notion developer tool.
[INDIRECT_PROMPT_INJECTION]: The skill defines an attack surface for indirect prompt injection because it reads and processes data from external Notion workspaces, which may contain malicious instructions.
Ingestion points: The skill uses ntn pages get, ntn datasources query, and ntn api v1/search to ingest data from pages and databases as described in SKILL.md.
Boundary markers: There are no instructions for the agent to use delimiters or to ignore embedded natural language instructions within the retrieved content.
Capability inventory: The skill allows for significant workspace modifications, including creating, editing, and trashing pages, as well as uploading files and performing raw API calls.
Sanitization: The documentation does not specify any sanitization or validation of the content returned from Notion before it is processed by the agent.
[COMMAND_EXECUTION]: The skill operates by executing subcommands of the ntn binary to perform Notion operations, which is the intended purpose of the skill.

notion