The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructions direct the agent to execute shell commands to open generated lesson files for the user. While typically used for benign 'open' or 'cat' operations, this establishes a pattern of executing local commands based on generated file paths.
Evidence in SKILL.md: "If possible, open the lesson file for the user by running a CLI command."
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it is designed to ingest and act upon data from multiple files within the local directory that may contain untrusted content (e.g., resources found online or mission statements provided by the user).
Ingestion points: SKILL.md mandates reading from MISSION.md, RESOURCES.md, ./learning-records/*.md, and ./assets/* to inform its teaching behavior.
Boundary markers: The provided templates (MISSION-FORMAT.md, RESOURCES-FORMAT.md, LEARNING-RECORD-FORMAT.md) do not include delimiters or instructions to ignore embedded commands within these data sources.
Capability inventory: The skill has the capability to write files to the local system and execute CLI commands.
Sanitization: No sanitization or validation of the content read from the workspace files is specified before it is interpolated into the agent's context.

teach