The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill provides instructions for installing the paper CLI tool directly from the vendor's GitHub repository (https://github.com/lassejlv/paper-cli) using the cargo install command.
[REMOTE_CODE_EXECUTION]: The paper upgrade command facilitates the download and execution of a remote installer script from the vendor's repository to update the CLI binary. It also utilizes npx skills@latest to update the skill definition.
[COMMAND_EXECUTION]: The core functionality of the skill relies on executing shell commands via the paper CLI to interact with a local MCP server. This includes capabilities for tool discovery, schema inspection, and executing design mutations.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from external design files and live MCP tool outputs, which could contain instructions intended to influence the agent's behavior.
Ingestion points: Data retrieved through paper call, paper inspect, and paper tools (SKILL.md).
Boundary markers: None identified; there are no specific instructions for the agent to distinguish between design data and behavioral instructions.
Capability inventory: Subprocess execution via shell commands, local file writes via the --output flag, and network communication via the MCP bridge (SKILL.md).
Sanitization: Tool arguments are required to be structured as JSON, but there is no specific sanitization for the content of the data being processed.

use-paper-cli