The Agent Skills Directory

[DATA_EXPOSURE_AND_EXFILTRATION]: The skill performs network operations exclusively to the official Linear GraphQL endpoint (https://api.linear.app/graphql). It facilitates authentication through Linear personal API keys or Zapier-managed OAuth connections. The documentation correctly advises users to manage credentials using environment variables or secret stores rather than hardcoding them, which aligns with security best practices.
[INDIRECT_PROMPT_INJECTION]: The skill includes an attack surface for indirect prompt injection because it ingests untrusted data from the Linear platform (such as issue titles, descriptions, and comments) into the agent's context. However, this is inherent to its primary purpose as an issue-tracking connector.
Ingestion points: scripts/getIssue.ts, scripts/searchIssues.ts, scripts/listIssueComments.ts, scripts/getProject.ts, and scripts/listProjects.ts.
Boundary markers: The skill uses a structured { data, meta } envelope for all tool outputs, providing clear separation between data and metadata.
Capability inventory: The skill has capabilities to create and modify resources in Linear, including createIssue, updateIssue, archiveIssue, createComment, createProject, and updateProject.
Sanitization: All inputs and outputs are strictly validated using Zod schemas, and the documentation in SKILL.md provides specific instructions for the agent on how to handle ambiguous data matches and ties to prevent accidental mis-execution.
[DYNAMIC_EXECUTION]: The cli.js entry point uses spawnSync to execute local TypeScript or JavaScript files within the skill's own directory structure. This is a standard dispatch mechanism for the platform and does not involve executing untrusted or remote code.

linear