The Agent Skills Directory

[COMMAND_EXECUTION]: The skill implements a workflow that involves creating a directory named _workspace/ at the project root and writing multiple markdown files to it, including the initial input (00_input.md) and subsequent analysis reports (e.g., 01_trend_analysis.md).
[PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection by ingesting and processing untrusted user data like exam results, goals, and study history.
Ingestion points: User inputs regarding exam names, subjects, goals, and existing performance data which are saved to _workspace/00_input.md for processing by various agent roles.
Boundary markers: The skill lacks explicit boundary markers or instructions to the agent to ignore potentially malicious instructions embedded within the user-provided content.
Capability inventory: The skill has the capability to write to the local file system and execute logical workflows based on the processed data.
Sanitization: There is no evidence of input validation or sanitization of the provided educational materials before they are processed by the agent team.

assessment-creator