The Agent Skills Directory

[COMMAND_EXECUTION]: The skill performs several automated shell operations to maintain the project environment.
Uses git extensively in Check 1 to manage stashing, branch switching, and fast-forwarding the main branch.
Executes npx skills update --project --yes in Check 2 to refresh the agent's installed capabilities.
[EXTERNAL_DOWNLOADS]: The skill initiates remote downloads to keep components updated.
The npx command in Check 2 fetches and executes the latest version of the skills package from the npm registry.
[PROMPT_INJECTION]: The skill is subject to indirect prompt injection risks due to its integration with external task management software.
Ingestion points: Untrusted data enters the agent context via the Linear API when fetching open tickets and backlog items in Check 3.
Boundary markers: None are present; there are no instructions to the agent to treat ticket content as data only or to ignore embedded instructions within ticket descriptions.
Capability inventory: The skill has the capacity to modify the local filesystem (via git and skill updates) and initiates the ship skill to perform further code modifications.
Sanitization: There is no mention of filtering, escaping, or validating the content of Linear tickets before the agent evaluates them to decide on the next 'ship' action.

project-health-check