The Agent Skills Directory

[PROMPT_INJECTION]: The skill facilitates an indirect prompt injection surface by passing user-supplied feedback directly into the prompts of sub-agents such as ui-ux-specialist and tech-lead-specialist.
Ingestion points: User feedback is ingested via command-line arguments or the AskUserQuestion tool in SKILL.md.
Boundary markers: The skill uses triple dashes (---) to wrap the {user_feedback} variable in the sub-agent task descriptions.
Capability inventory: The involved sub-agents have the ability to read from and write to markdown and JSON files within the project repository using the Write tool.
Sanitization: There is no evidence of input validation, escaping, or explicit instructions for the sub-agents to disregard instructions embedded within the user feedback, allowing a malicious user to potentially manipulate sub-agent behavior.

review-plan