The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from Linear tickets and documents which could potentially influence agent behavior if malicious instructions are embedded.
Ingestion points: The skill is instructed to read Linear issue descriptions via the get_issue tool, Linear project documents via the get_document tool, and local project-specific configuration files such as AGENTS.md and CLAUDE.md.
Boundary markers: No explicit boundary markers, delimiters, or instructions to treat ticket-sourced content as untrusted are present in the skill definition.
Capability inventory: The skill possesses the ability to read local source code, perform Linear API write operations (such as update_issue to change status and create_comment to add evidence), and execute local git operations (creating branches via git branch).
Sanitization: The skill lacks mechanisms for sanitizing or validating the content retrieved from external Linear sources before using it to load project context or inform the agent's workflow steps.

kata-linear