AI agents call search_context to retrieve information from Codetex without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This is a read-only operation that queries pre-indexed code context via semantic search. It retrieves data without side effects, matching the Read category definition. Severity is low because even if an AI misuses this tool by requesting sensitive code sections, the worst outcome is information disclosure of code already in the indexed repository, not execution, modification, or deletion.
From the tool's definition Tool name is 'search_context' and description states it 'Search[es] for relevant code context using semantic similarity.' The action is querying a vector database (SQLite vector search as mentioned in server description) to retrieve code context.
Attacks that exploit this kind of access
Search for relevant code context using semantic similarity. It is categorised as a Read tool in the Codetex MCP Server, which means it retrieves data without modifying state.
Register the Codetex MCP server in PolicyLayer and add a rule for search_context: 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 Codetex. Nothing to install.
search_context 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_context 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_context. 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_context is provided by the Codetex MCP server (mrosata/codetex-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 →