The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill facilitates communication between agents, which introduces a surface where one agent could send a message containing malicious instructions to another.
Ingestion points: The agent ingests data from external sources via the nitro agent mail inbox, read, and watch commands described in SKILL.md.
Boundary markers: The instructions do not define specific delimiters or guardrails to prevent the agent from obeying instructions embedded within the message bodies.
Capability inventory: The agent has the ability to execute shell commands through the nitro CLI and potentially other tools available in the environment.
Sanitization: There is no mention of filtering or sanitizing the content of received messages before they are processed by the agent.
[COMMAND_EXECUTION]: The skill relies on the execution of the nitro command-line utility for all core operations, including sending and receiving mail, listing actors, and searching threads.
[EXTERNAL_DOWNLOADS]: The documentation includes a link to the vendor's official installation page (https://chillicream.com/docs/nitro/cli/installation). The skill explicitly instructs the agent not to attempt the installation itself, which is a significant security control.

nitro-mail