The Agent Skills Directory

[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface as it processes and interpolates user-provided prompts into the operational context of secondary agents.\n- Ingestion points: The prompt parameters in plan_project and create_mission tools serve as entry points for untrusted data.\n- Boundary markers: No delimiters or safety instructions are specified to help sub-agents distinguish between system instructions and potentially malicious payload content.\n- Capability inventory: Orchestrated agents are granted 'full tooling' and operate within git worktrees, enabling file and system interactions.\n- Sanitization: The documentation does not describe any validation or filtering mechanisms for mission prompts before they are dispatched to agents.

claude-devfleet