Operating Multica

Your runtime brief owns the per-turn workflow: which issue you are on, when to comment, what status to write. This skill owns the platform contracts behind it, what a command actually does, what the server validates, and which writes have consequences you cannot take back.

Read the invariants below, then open the reference(s) your task actually needs, usually one, sometimes a few. Do not read them all.

Routing

Open	When the task is about
references/issues.md	Issues: PR linking, reading a linked PR's state, custom properties, status side effects, sub-issues and stages, who else is running, wakeups (events, conditions, timers, check-ins), charts vs attached files in a comment
references/mentions.md	Writing a mention:// link: which types enqueue a run, which are inert, why one silently did nothing; steering another agent's running turn (comment add --steer)
references/agents.md	Creating, copying or debugging an agent definition: fields, secrets, MCP config, skill binding
references/squads.md	Squads: leader routing, roster, recording leader activity, why a squad did or did not run
references/autopilots.md	Autopilots: schedule / webhook / manual triggers, create_issue vs run_only, why one did not fire
references/projects.md	Projects and their durable resources (github_repo, local_directory, worktree mode); project workflows, status columns, handoffs, a [Project workflow] block in your prompt, reading or changing a workflow (multica workflow, multica project workflow)
references/runtimes.md	Runtimes, daemons, repo checkout, and the task CLI boundary
references/skill-import.md	Importing a skill into this workspace from a URL or a local archive