Search for code by name or meaning. Uses vector retrieval + cross-encoder reranking.
AI agents call search_code to retrieve information from CodeGraph without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This is a pure retrieval and query tool with no side effects. It searches and returns code information without creating, modifying, deleting, or executing anything. The mechanism (vector retrieval + reranking) confirms it is a read-only search utility. Even if an AI agent misuses this tool, the worst outcome is accessing code that shouldn't be read—a confidentiality issue but not destructive or executable.
From the tool's definition Tool description states it 'Search[es] for code by name or meaning' using 'vector retrieval + cross-encoder reranking'. The server description emphasizes 'search code', 'explore codebase structure', and 'read source code' as its core functions.
Documented attack patterns abuse exactly the kind of access search_code gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and CodeGraph, and nothing reaches the server without passing your rules. This is the rule we recommend for search_code:
{
"version": "1",
"default": "deny",
"tools": {
"search_code": {}
}
}search_code is read-only, so it stays allowed — but everything else on the server is denied unless you say otherwise.
Free to start. No card required.
Search for code by name or meaning. Uses vector retrieval + cross-encoder reranking. It is categorised as a Read tool in the CodeGraph MCP Server, which means it retrieves data without modifying state.
Register the CodeGraph MCP server in PolicyLayer and add a rule for search_code: 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 CodeGraph. Nothing to install.
search_code is a Read tool with low risk. Read-only tools are generally safe to allow by default.
Yes. Add a rate_limit block to the search_code 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 search_code. 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.
search_code is provided by the CodeGraph MCP server (phoenixrr2113/codebase-graph). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from CodeGraph, add the rest of your stack, and see everything your agents can call. Then put policy on all of it.
Free to start. No card required.
18 CodeGraph tools catalogued and risk-classified — across an index of 43,000+ MCP servers.