The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to its data ingestion patterns and available tool capabilities.
Ingestion points: The skill pulls untrusted data from external sources using Linear MCP tools such as get_issue and list_issues (defined in Path 2 of SKILL.md).
Boundary markers: There are no explicit boundary markers or instructions provided to the agent to disregard instructions embedded within the fetched Linear data.
Capability inventory: The skill has write access to the external platform, allowing it to execute create_comment and create_issue actions based on the processed (and potentially malicious) input.
Sanitization: The skill does not define any sanitization, validation, or escaping protocols for the external content before it is interpolated into the agent's context.

context-engineering