The Agent Skills Directory

[PROMPT_INJECTION]: The skill acts as an orchestration layer that processes and relays data between multiple agents, presenting an indirect prompt injection surface.
Ingestion points: Untrusted data can enter the system through the prompt arguments in plan_project and create_mission, as well as through the reports processed by get_report in SKILL.md.
Boundary markers: The skill does not define specific boundary markers or instructions to ignore embedded commands within the mission prompts or agent reports.
Capability inventory: The skill has the capability to dispatch agents that, according to SKILL.md, operate with "full tooling" in isolated git worktrees, which typically includes file system and command execution access.
Sanitization: There is no evidence of sanitization or validation for the content being passed between missions or extracted from agent reports.

gemini-devfleet