Branch Coordination for AI Agents
Your AI team's coordination layer. Track branches, detect conflicts early, and get optimal merge order — so your agents can work in harmony.
npm install -g spidersan
Detect file conflicts between branches before you go too deep. Know exactly who's touching what files.
Daemon mode that auto-registers file changes as you work. Real-time conflict warnings with Hub integration.
Get topologically-sorted merge order. Know exactly which branches to merge first for zero conflicts.
Got 10+ abandoned branches? Scan, triage (MERGE/SALVAGE/ABANDON), and salvage good code from broken branches.
Show git history combined with agent activity for any files. See the full story of who changed what.
Full Model Context Protocol server for AI agents. Seamless integration with Claude, Cursor, and other AI tools.
Pause before merging — request human approval via Toak when Spidersan detects a conflict that needs a judgment call.
spidersan ready-check --require-approval
Store your Spidersan Hub API keys securely in envoak. Share across machines without exposing secrets.
envoak push # includes SPIDERSAN_HUB_KEY
Query branch history and conflict patterns from the knowledge base. Learn from past coordination challenges.
watsan oracle "common auth conflicts"
spidersan initspidersan register --filesspidersan listspidersan conflictsspidersan merge-orderspidersan ready-checkspidersan dependsspidersan stalespidersan cleanupspidersan rescuespidersan abandonspidersan mergedspidersan syncspidersan watchspidersan doctorspidersan configspidersan autospidersan welcome
Advanced commands like who-touched,
monitor,
lock, and
scan
are available in the ecosystem plugin.
Running 5 Claude agents in parallel? Spidersan ensures they don't step on each other's toes with real-time conflict detection.
Got 20 branches from a weekend sprint? Use rescue mode to scan, triage, and salvage the good parts automatically.
Human + AI working together? Both register their branches and get instant alerts when overlapping work is detected.
Orchestrating a flock of specialized agents? Spidersan provides the coordination layer that prevents chaos.
Spidersan is a coordination layer for AI coding agents. It tracks which branches are modifying which files, detects potential merge conflicts before they happen, and provides optimal merge ordering so your agents can work in parallel without stepping on each other.
When you register a branch with spidersan register --files, Spidersan tracks
which files you're modifying. When you run spidersan conflicts, it compares
your file list against all other registered branches and alerts you to any overlaps. In
watch mode, this happens automatically in real-time.
Yes! All 18 core CLI commands are MIT licensed and free forever — including init, register, conflicts, merge-order, watch, doctor, rescue, and more. Advanced ecosystem commands are available via the ecosystem plugin.
Run spidersan watch --agent myagent to start watching. Add --hub
for real-time conflict warnings via the Spidersan Hub. Use --quiet to only show
conflicts and suppress normal output. The watcher auto-registers files as you edit them with
1-second debouncing.
Spidersan works with any git-based workflow. It has explicit MCP integration for Claude, Cursor, and other AI tools that support Model Context Protocol. For other tools, use the CLI directly or integrate via the watch mode daemon.