linear
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to read external data from Linear (issues, comments, and documents) which can contain instructions intended to influence the agent's behavior.
- Ingestion points: The skill uses methods such as
issues.readanddocuments.readto fetch text-heavy content from the Linear workspace into the agent's context. - Boundary markers: The instructions do not specify the use of delimiters or 'ignore' instructions to isolate the external content from the agent's system prompt.
- Capability inventory: The agent possesses write capabilities within Linear (e.g.,
issues.manage,projects.create) and potentially access to other tools in the execution environment. - Sanitization: There is no evidence of content sanitization, filtering, or validation before the retrieved data is interpolated into the prompt.
- [COMMAND_EXECUTION]: The skill operates by executing the
linearCLI binary for all actions, including authentication and tool calls. This is the intended functional mechanism of the skill.
Audit Metadata