The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection via the task files it processes. * Ingestion points: Task content is ingested via the npx @jagersoftware/fine show command. * Boundary markers: There are no instructions to treat task content as untrusted data or use delimiters. * Capability inventory: The agent is granted the ability to write code, create files, and execute arbitrary shell commands. * Sanitization: There is no evidence of sanitization or validation of the task steps before implementation.
[EXTERNAL_DOWNLOADS]: The skill uses npx to download and execute the @jagersoftware/fine package. This resource is associated with the vendor 'quinlanjager'.
[COMMAND_EXECUTION]: The workflow explicitly directs the agent to "run commands, whatever the step requires" and execute tests using bun test, which allows for arbitrary code execution driven by external task data.

complete-task