AI agents use gh_pr_create to create or update resources in Gh — usually the action step of a workflow, after the agent has gathered context. Every call changes real data in your Gh environment.
This tool creates pull requests, which is a reversible write operation that modifies repository state by adding new data. It has moderate blast radius—a malicious agent could spam PRs, modify workflows, or introduce malicious code into review processes—but the action is not destructive (PRs can be closed/deleted) and does not move money or execute arbitrary code directly.
From the tool's definition Tool name 'gh_pr_create' indicates creation of pull requests. Server description states it enables 'creation and editing of issues and pull requests.' The tool operates within GitHub CLI operations for creating/editing PR data.
Attacks that exploit this kind of access
gh_pr_create. It is categorised as a Write tool in the Gh MCP Server, which means it can create or modify data. Consider rate limits to prevent runaway writes.
Register the Gh MCP server in PolicyLayer and add a rule for gh_pr_create: 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 Gh. Nothing to install.
gh_pr_create 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 gh_pr_create 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 gh_pr_create. 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.
gh_pr_create is provided by the Gh MCP server (yukkie/gh-mcp). 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 →