The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external sources such as Linear issues, projects, documents, and comments, which are then interpolated into the agent's context. This data could contain malicious instructions designed to influence the agent's subsequent actions. 1. Ingestion points: lin issue get, lin project get, lin initiative get, lin document get, and lin issue comment list (SKILL.md). 2. Boundary markers: The skill documentation notes that long text fields are truncated to 200 characters by default (references/output.md), but this can be bypassed using the --full or --expand flags. 3. Capability inventory: The skill has broad execution capabilities via Bash(lin *), including the ability to modify or delete resources and execute raw GraphQL queries. 4. Sanitization: No sanitization of the ingested content is mentioned.
[EXTERNAL_DOWNLOADS]: The skill allows downloading files from Linear's official CDN. Evidence: The lin file download command (SKILL.md) targets https://uploads.linear.app/.
[COMMAND_EXECUTION]: The skill provides an interface for executing Linear operations. Evidence: The lin tool is invoked through the Bash tool with broad arguments (lin *), as seen in the allowed-tools section of SKILL.md.
[DYNAMIC_EXECUTION]: The skill includes an 'escape hatch' for executing raw GraphQL queries. Evidence: The lin api query command (SKILL.md) allows the execution of arbitrary GraphQL strings against the Linear API.

lin