The Agent Skills Directory

[COMMAND_EXECUTION]: The skill allows the agent to spawn new processes and execute arbitrary shell commands in managed terminals.
Evidence: The commands orca terminal create, orca terminal send, and orca orchestration run provide the ability to start agent CLIs and deliver instructions to them directly.
[PROMPT_INJECTION]: The skill provides an attack surface for indirect prompt injection by facilitating communication between multiple agents and utilizing a preamble injection feature.
Ingestion points: Data from other agents enters the context through orca orchestration check, orca orchestration inbox, and orca terminal read in the SKILL.md file.
Boundary markers: The documentation does not specify the use of delimiters or instructions to ignore embedded commands when processing messages from other agents.
Capability inventory: The coordinator possesses significant capabilities, including terminal creation and the execution of shell commands across the environment.
Sanitization: There is no evidence of sanitization or validation of message bodies or task results before they are processed by the coordinator or injected into the prompts of other agents via the dispatch --inject command.

orchestration