The Agent Skills Directory

[SAFE]: The skill's primary function is the batch processing of Linear issues, which is performed through authorized MCP tool integrations as part of a multi-phase planning workflow.
[DATA_EXFILTRATION]: Analysis confirms that data handling is restricted to reading from the local state and writing to authorized Linear API endpoints or local audit logs. No unauthorized network activity or credential harvesting was detected.
[COMMAND_EXECUTION]: Use of the Bash tool is limited to benign path resolution tasks, such as utilizing git rev-parse to determine the project root directory.
[PROMPT_INJECTION]: The skill contains instructions for "gate-respect," which are designed to ensure the model adheres strictly to user choices made during interactive prompts. This is a reliability feature rather than a malicious attempt to bypass safety filters.
[INDIRECT_PROMPT_INJECTION]: The skill handles potentially untrusted data from earlier workflow phases (e.g., free-text answers). It implements proactive security measures by sanitizing and escaping this content before rendering it in markdown previews, reducing the risk of indirect injection attacks.

linear-housekeeping