Coordinate spawned agents across verification rounds until exit criteria are met.
npx skills add https://github.com/humblemuzzu/ghosttyyy --skill rounds
Please help me install this Agent Skill. Skill: rounds Source: https://github.com/humblemuzzu/ghosttyyy/tree/main/pi-setup/config-skills/rounds Command: npx skills add https://github.com/humblemuzzu/ghosttyyy --skill rounds
Coordinating spawned agents to verify results converge across rounds and ensure decisions are validated against explicit criteria.
Initiate a multi-round verification by spawning agents, collecting reports, and escalating fixes until two consecutive clean rounds are achieved.
High-intent search queries and answers about installing and using this skill.
You coordinate multi-agent verification by orchestrating a round lifecycle that spawns agents, collects reports, assesses convergence, and escalates fixes until explicit exit criteria are satisfied across consecutive clean rounds.
The round lifecycle for multi-round verification follows a spawn, wait, assess, and escalate sequence. It tracks issues and advances iterations until decisions validate against defined exit criteria and achieve two consecutive clean rounds.
Exit criteria for verification rounds are defined explicitly to determine when iterations are considered clean. The process documents changes across iterations and requires two consecutive clean rounds to achieve convergence.
Yes, inter-agent coordination requires prerequisite skills like spawn, coordinate, and report. These skills handle sub-tasks and enable the orchestration of the round lifecycle to verify results effectively.
Yes, you can orchestrate parallel debates and handle sub-tasks through composability with spar. This allows dedicated skills to manage sub-tasks while coordinating spawned agents across verification rounds.
Multi-round verification applies to bounded tasks requiring explicit exit criteria and issue tracking. You should avoid this approach for unbounded tasks or processes that lack strict convergence validation requirements across iterations.
memorysaverConstruct and manage dynamic multi-agent workflows in Claude Code.
yannicklothAutomate iterative review-fix loops until convergence or 12 rounds.
edwifiguyDelegate complex subtasks to sub-agents with self-contained context.
maystudiosCoordinate parallel agent execution with git worktree isolation and PASS/FAIL results.
