The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes data generated by external agents in a multi-agent swarm.
Ingestion points: The skill retrieves external data using tools like read-messages, get-tasks, get-task-details, and poll-task in SKILL.md and MCP-REFERENCE.md.
Boundary markers: There are no instructions for the agent to treat incoming message content or task descriptions as untrusted data or to use specific delimiters.
Capability inventory: The skill possesses significant capabilities, including task delegation (send-task), service registration (register-service), and agent communication (post-message).
Sanitization: No sanitization or validation protocols are described for handling the contents of messages or tasks received from other swarm participants.
[SAFE]: The skill references technical documentation hosted on GitHub at github.com/desplega-ai/agent-swarm. This reference belongs to the skill author's own organization and represents legitimate project documentation.

swarm-expert