The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external documentation files which could contain malicious instructions designed to subvert the agent's logic during the requirement validation process.
Ingestion points: The skill ingests untrusted data from multiple files including spec.md, ux.md, ui.md, plan.md, tasks.md, data-model.md, contracts/, research.md, and setup.md within a user-provided feature path.
Boundary markers: There are no explicit boundary markers or instructions defined to isolate the content of these files from the agent's system instructions or to ignore embedded commands.
Capability inventory: The skill possesses significant capabilities including the Read and Write tools (allowing it to modify the project's tasks.md) and the Bash tool (used for directory creation and file verification).
Sanitization: The skill lacks sanitization logic to filter or escape potentially malicious natural language instructions found within the ingested documentation artifacts.

validation