Code-aware local generation. Claude reads source code with Read/Grep (free), then delegates the TEXT GENERATION to local - docstrings, test stubs, explanations, type annotations, inline comments, or review feedback. This is the primary tool for reducing API token usage on code tasks. Claude orche...
AI agents call local_code to retrieve information from Mcp Ollama without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
This tool takes code as input and produces generated text (docstrings, explanations, stubs, annotations, etc.) using a local Ollama model. It is a read/generation tool with no side effects: it reads source code and outputs text. The orchestrating agent (Claude) must separately decide to write any output back to files.
From the tool's definition 'delegates the TEXT GENERATION to local', 'docstrings, test stubs, explanations, type annotations, inline comments, or review feedback' — the tool generates text output only; it does not write, execute, or modify files
Attacks that exploit this kind of access
Code-aware local generation. Claude reads source code with Read/Grep (free), then delegates the TEXT GENERATION to local - docstrings, test stubs, explanations, type annotations, inline comments, or review feedback. This is the primary tool for reducing API token usage on code tasks. Claude orchestrates (decides what code to read, what task to perform), but the actual generation happens locally at zero API cost. Accepts up to ~12K tokens of code context (16K model context minus overhead). For larger contexts, break into focused chunks (one function, one class). It is categorised as a Read tool in the Mcp Ollama MCP Server, which means it retrieves data without modifying state.
Register the Mcp Ollama MCP server in PolicyLayer and add a rule for local_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 Mcp Ollama. Nothing to install.
local_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 local_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 local_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.
local_code is provided by the Mcp Ollama MCP server (true-alter/mcp-ollama). 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 →