The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the Letta CLI (letta) to send and manage messages between agents. This is the core intended functionality and uses the vendor's official command-line tools.
[PROMPT_INJECTION]: Evaluated the messaging surface for indirect prompt injection vulnerabilities.
Ingestion points: The message text parameter in the letta CLI commands within SKILL.md.
Boundary markers: The documentation explains that a <system-reminder> block is automatically injected to provide context and identity to the receiving agent.
Capability inventory: The skill is strictly scoped to communication and does not have permissions to access the local file system or execute arbitrary shell commands.
Sanitization: No explicit sanitization is described in the skill instructions; it relies on the platform's native agent communication security.
[SAFE]: No findings related to hardcoded credentials, unauthorized data exfiltration, or malicious persistence mechanisms were detected. The use of the $LETTA_AGENT_ID environment variable is a safe practice for agent identification.

messaging-agents