AI agents call codegen_assemble to retrieve information from Pen without modifying anything — typically the context-gathering step in research, monitoring, and reporting workflows, before the agent takes action elsewhere.
Despite the side effect of clearing internal cache, this tool's core function is to retrieve and return data (chunk results) in a specific order. It does not create, modify, delete, or overwrite user data irreversibly, nor does it execute arbitrary code or move money.
From the tool's definition Tool description states 'Retrieve all chunk results in topological order for final assembly' and 'Returns degraded=true if any chunk is not done.' The primary action is retrieval/querying of existing chunk results.
Attacks that exploit this kind of access
Retrieve all chunk results in topological order for final assembly. Terminal operation — clears plan cache after returning. Returns degraded=true if any chunk is not done. It is categorised as a Read tool in the Pen MCP Server, which means it retrieves data without modifying state.
Register the Pen MCP server in PolicyLayer and add a rule for codegen_assemble: 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 Pen. Nothing to install.
codegen_assemble 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 codegen_assemble 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 codegen_assemble. 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.
codegen_assemble is provided by the Pen MCP server (@zseven-w/pen-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 →