Rosentic checks whether your active branches are compatible with each other before merge. It detects broken function signatures, HTTP route conflicts, and schema mismatches across 15+ languages using deterministic AST analysis.
This is not AI code review. Rosentic finds structural contract conflicts that would cause runtime failures after merge - the kind of breaks that tests don't catch because they only run on one branch at a time.
pip install rosentic-mcpThen add to your editor's MCP config:
{
"mcpServers": {
"rosentic": {
"command": "rosentic-mcp",
"args": [],
"env": {}
}
}
}Works with Cursor, Claude Code, Cline, Windsurf, and any editor that supports MCP.
Rosentic gives your AI coding agent three MCP tools:
| Tool | Purpose |
|---|---|
check_conflicts |
Scan a repo for cross-branch contract conflicts |
explain_conflict |
Explain a specific finding and what to do about it |
list_branches |
List active branches so the agent can choose what to scan |
- Branch A adds a required parameter to
createOrder(). Branch B callscreateOrder()with the old signature. Both pass CI. Merge breaks production. - Branch A changes
POST /api/usersto requireemailin the request body. Branch B's frontend still sends the old payload. - Branch A removes a field from a GraphQL type that Branch B's query depends on.
Before you push, open a PR, or merge, your agent runs a cross-branch compatibility check. It calls check_conflicts, which scans your repo's branch heads using tree-sitter AST parsing and reports any conflicts.
No code leaves your machine. The scan runs locally against your git repo.
Semgrep Guardian checks each file for security vulnerabilities and code issues. Rosentic checks whether your branches are compatible with each other. They solve different problems and work well together:
- Semgrep Guardian: "Is this code safe and correct on its own?"
- Rosentic: "Will this branch break other branches when merged?"