Create a new branch in a GitHub repository
AI agents use create_branch to create or update resources in Mcp Github — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Mcp Github environment.
Creating a branch is a write operation that modifies repository state reversibly. While it has moderate blast radius in development workflows (could create unwanted branches, clutter, or interfere with CI/CD), it is not destructive (branches can be deleted), does not execute code, and does not involve financial transactions.
From the tool's definition Tool description states 'Create a new branch in a GitHub repository' — this creates new data (a branch) in a reversible manner. Creating a branch does not permanently destroy data or execute arbitrary code, but it does modify the repository structure.
Attacks that exploit this kind of access
Create a new branch in a GitHub repository. It is categorised as a Write tool in the Mcp Github MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the Mcp Github MCP server in PolicyLayer and add a rule for create_branch: allow, deny, rate-limit, or require approval. Point your MCP client at the PolicyLayer proxy URL and the rule is enforced on every call, before it reaches Mcp Github. Nothing to install.
create_branch is a Write tool with medium risk. Write tools should be rate-limited to prevent accidental bulk modifications.
Yes. Add a rate_limit block to the create_branch rule in your PolicyLayer policy. For example, setting max: 10 and window: 60 limits the tool to 10 calls per minute. Rate limits are tracked per agent session and reset automatically.
Set action: deny in the PolicyLayer policy for create_branch. The AI agent will receive a policy violation error and cannot call the tool. You can also include a reason field to explain why the tool is blocked.
create_branch is provided by the Mcp Github MCP server (@missionsquad/mcp-github). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Every MCP server has a record like this.
Type a name, get the same breakdown: verified identity, auth posture, risk grade, capabilities, recommended policy.
Teams ship this data inside their own products. See what a licence covers →