Prevent file conflicts between multiple AI agents working in parallel
/plugin marketplace add ekson73/multi-agent-os/plugin install multi-agent-os@multi-agent-osThis skill inherits all available tools. When active, it can use any tool Claude has access to.
Prevent file conflicts between multiple AI agents working in parallel. Implements a 7-phase workflow with mandatory worktree usage, lock files, and QA validation.
git worktree list # Check active worktrees
cat .worktrees/tasks.md # Check IN_PROGRESS tasks
cat .worktrees/sessions.json # Check active sessions
ls .worktrees/*.lock # Check lock files
git status # Check uncommitted changes
git log --oneline -3 # Verify expected state
git worktree add .worktrees/{agent-hex}-{feature} -b {tipo}/{name}
cd .worktrees/{agent-hex}-{feature}
For protected files (CLAUDE.md, README.md, CSVs):
cp .worktrees/session_lock.template.json .worktrees/{id}.lockgit status --short | grep {file}
# M or MM → DO NOT EDIT (another agent working)
# ?? or CLEAN → Safe to edit
High-risk files requiring explicit coordination:
git add {specific-files} (NEVER git add .)tipo(escopo): desc - Agent: {name}For sessions >1 hour:
git fetch origin
git log HEAD..origin/main --oneline
Before closing session:
Skill based on Anti-Conflict Protocol v3.2 | multi-agent-os