The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest data from a Linear workspace (such as issues, comments, and documentation) and use that data to perform further actions. This creates a surface for indirect prompt injection where an attacker with access to the Linear workspace could embed instructions in ticket descriptions or comments to manipulate the agent's behavior.
Ingestion points: The skill uses tools like list_issues, get_issue, search_documentation, and list_comments in SKILL.md to read external data.
Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present in the processing logic.
Capability inventory: The skill has write capabilities including create_issue, update_issue, create_project, and create_comment in SKILL.md.
Sanitization: No sanitization or validation of the ingested content is specified.
[EXTERNAL_DOWNLOADS]: The skill guides the user or agent to fetch configuration and executable code from external sources to enable the Linear integration.
Evidence: References to https://mcp.linear.app/mcp and https://mcp.linear.app/sse in SKILL.md for MCP server configuration.
[COMMAND_EXECUTION]: The instructions include commands that modify the local environment and execute external packages.
Evidence: Shell commands in SKILL.md such as codex mcp add linear, codex mcp login linear, and the use of npx -y mcp-remote to run a remote package.

linear