The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed as the first stage of a multi-step chain where its output (ideator-output/project-brief.md) is used as an instruction source for downstream agents such as constructor and datasecurer.
Ingestion points: The skill collects free-form user answers to Socratic questioning during Phase 1, 2, and 3 of its workflow in SKILL.md.
Boundary markers: The templates/project-brief.md uses markdown headers but lacks explicit delimiters or instructions to prevent downstream agents from executing commands potentially embedded in the user-provided project description.
Capability inventory: The skill utilizes the write and edit tools to create the handoff artifacts.
Sanitization: The skill does not perform sanitization, validation, or escaping of user input before interpolating it into the final project brief, creating a surface for handoff poisoning.

ideator