The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill aggregates user-supplied text into a founder/board/brief.md file which is then used as instructions for sub-agents, creating a surface for indirect prompt injection.
Ingestion points: User input from $ARGUMENTS or conversation history is used to create founder/idea.md and founder/board/brief.md as described in SKILL.md.
Boundary markers: The skill does not employ delimiters (e.g., XML tags) to isolate user-provided data from the agent's instructions in the briefing file, increasing the risk that embedded commands are treated as instructions.
Capability inventory: Sub-agents are launched in SKILL.md to read project files and write analysis memos to the filesystem (founder/board/).
Sanitization: There is no evidence of input validation or instruction filtering to prevent malicious payloads from being passed to the sub-agents.

founder-board