The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute the ajv command-line tool using shell interpolation. Specifically, the {type} variable, which is extracted from the meta.type field of a loaded JSON file, is used to construct the file path for the validation schema.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection attack surface by processing untrusted data from local files.
Ingestion points: The skill reads a JSON file from a path determined by the user-provided session-slug input at docs/instructional-design/session-plans/{slug}/session-plan.json.
Boundary markers: There are no instructions or delimiters provided to the agent to treat the contents of the JSON file as untrusted data or to ignore any embedded instructions.
Capability inventory: The agent uses shell command execution (ajv validate) and internal logic (regex matching and arithmetic checks) to process the data.
Sanitization: The instructions lack explicit requirements for the agent to sanitize or validate the meta.type field before incorporating it into the shell command string, creating a risk of command injection.

session-plan-audit