The Agent Skills Directory

[Indirect Prompt Injection] (LOW): The skill is designed to handle and act upon data provided by other agents, which creates a significant surface for indirect prompt injection.
Ingestion points: The agent ingests untrusted data through tools such as read-messages, get-tasks, get-task-details, and poll-task as described in MCP-REFERENCE.md.
Boundary markers: The skill documentation does not provide instructions for using delimiters or boundary markers to distinguish agent-provided data from system-level instructions.
Capability inventory: The skill possesses capabilities to create and assign tasks (send-task, task-action), communicate in channels (post-message), and register services (register-service).
Sanitization: There are no documented procedures for sanitizing or validating the content of messages or task descriptions before they are processed.
[No Code] (SAFE): The provided files consist solely of markdown documentation and reference information; no scripts, binaries, or automated installation processes are included in the skill.

swarm-expert