The Agent Skills Directory

[PROMPT_INJECTION]: The skill is designed to read and process content from a Linear workspace, which serves as an ingestion point for untrusted data. (1) Ingestion points: Issue descriptions, project data, and comments fetched via tool calls like get_issue or list_comments (documented in SKILL.md). (2) Boundary markers: There are no specific instructions for the agent to use delimiters or ignore embedded instructions when processing workspace data. (3) Capability inventory: The skill includes tools to create or modify issues, labels, and comments, providing a path for potential multi-step injection effects. (4) Sanitization: No content validation or sanitization steps are defined for data retrieved from the workspace.
[EXTERNAL_DOWNLOADS]: Instructions in references/setup.md guide the user to use npx to fetch and run the mcp-remote package, enabling connection to the official Linear MCP endpoint.
[COMMAND_EXECUTION]: The skill documentation includes setup commands for the Codex environment and configuration snippets for Windows/WSL that utilize shell commands for integration purposes.

linear