Menu
SkillsMP has collected 9 skills from fade/dsmr-mcp. Open a skill to review its source and details.
Showing 9 of 9 collected skills.
Rebuild a leader's repo-local context AND the whole-fleet context in one pass, so the leader can assert positive, rapid control over every worker on ITS OWN bus. Use when you are a leader starting or resuming a session, after a fleet restart, when you have…
Read a PR body, commit message, diff comments, or any artifact about to be placed in front of the operator, as a person who knows the codebase and did not attend the session that produced it. Asks first whether the text reads as its author writing about his…
Rebuild a worker (sister) agent's repo-local context, announce to the leader in one line, revalidate the leader/worker contract, PRE-CLEAR every permission the assigned work needs so a denial never blocks mid-task, and arm a watch on each joined bus while…
Stay reachable on the dsmr-mcp coordination bus without polling: arm ONE persistent Monitor over the streaming watcher per joined bus, and each new message wakes you. Use when you need to listen for bus messages in the background, wake on a bus event, or stay…
Coordinated take-down and bring-up of a multi-agent fleet (one leader plus N sister agents in other repos) coordinating over a durable bus. Use when the operator asks to restart the agents/fleet/constellation, when you are a leader or sister coming back up…
Assemble a fleet on its own named bus from the leader's terminal: validate the names, declare the leader, enroll each sister, hand every repo its tag, and print the arm line each sister will use. Use when starting a new fleet, adding or removing a participant…
Scaffold a new Common Lisp project from the dsmr-mcp skeleton when run in an empty directory. Generates a package-inferred ASDF system, packages, a zebra smoke test, and delivery-discipline files via the dsmr-mcp MCP server.
Retire a running dsmr-mcp bus broker and bring up a fresh one on current code, without losing messages or rotating the WAL. Use when the broker has been up long enough to be serving stale source, when a broker must be restarted to pick up a fix, or when the…
Snapshot in-flight state to a handoff file so the next instance picks up cleanly without re-prompting.