The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted repository content which presents a surface for indirect prompt injection.
Ingestion points: Reads instructions from AGENTS.md, CLAUDE.md, and design documentation from design/gdd/, design/quick-specs/, and Assets/Game/Data/ as specified in SKILL.md and references/claude-skill.md.
Boundary markers: Absent; the skill does not utilize specific delimiters or instructions to ignore embedded commands in the ingested files.
Capability inventory: Includes Write access to design/quick-specs/ and Edit access to design/gdd/ as defined in references/claude-skill.md.
Sanitization: Absent; the skill does not explicitly sanitize or validate the content of the ingested files before using it to generate or modify specifications.
Context: The risk is minimized as the skill is designed to facilitate developer workflows and mandates human approval via the AskUserQuestion tool before any changes are committed to the filesystem.

quick-design