The Agent Skills Directory

[COMMAND_EXECUTION]: The skill relies on shell commands to interact with the reMarkable cloud API and manage local files.
Evidence: Instructions include the use of rmapi for cloud operations, unzip for file extraction, pandoc for document conversion, and open for viewing files.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data (file names and document metadata) retrieved from the reMarkable tablet, which constitutes an attack surface for indirect prompt injection.
Ingestion points: File listings and metadata retrieved via rmapi ls and rmapi get (SKILL.md).
Boundary markers: None present; the skill does not instruct the agent to ignore instructions embedded in file names or metadata.
Capability inventory: The skill has the ability to execute shell commands, perform network requests via the rmapi tool, and read/write to the /tmp/ directory.
Sanitization: There is no evidence of sanitization or validation of the file names or content retrieved from the external device.

remarkable-expert