The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the multica command-line utility for managing autopilot lifecycle, including creation, updates, and triggering. The tool usage is constrained to the multica binary as defined in the allowed-tools frontmatter.
[DATA_EXFILTRATION]: Contains explicit instructions for the agent to redact webhook tokens and signing material, preventing accidental exposure of secrets in logs, comments, or documentation.
[PROMPT_INJECTION]: The skill manages the creation of autopilot tasks that incorporate user-provided descriptions. This represents an indirect prompt injection surface for the agents that eventually execute these tasks, although the management interface itself does not contain injection patterns.

multica-autopilots