The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the official ticketfairy command-line interface to perform administrative tasks on the platform.
Evidence: Commands such as ticketfairy event-os risks add, ticketfairy event-os incidents report, and ticketfairy login are used to interact with the event management system.
[EXTERNAL_DOWNLOADS]: The skill instructs the user to install an external package from a public registry.
Evidence: Recommends npm install -g ticketfairy to set up the necessary tooling. This package is the official tool provided by the vendor.
[INDIRECT_PROMPT_INJECTION]: The skill has a functional surface for indirect prompt injection as it processes external, untrusted documents like technical riders and supplier quotes to suggest budget or risk register updates.
Ingestion points: Reads technical riders, supplier quotes, and external incident reports (SKILL.md).
Boundary markers: The instructions explicitly state to follow documented guides and always wait for the organizer's approval before committing any changes.
Capability inventory: The agent can modify sensitive event data (risks, budgets, contracts) via CLI and MCP tools (manage_event_risks, send_contract_for_signature).
Sanitization: Relies on a human-in-the-loop workflow where the agent must show the proposed change and receive explicit confirmation before execution.

run-the-site-and-safety