CODEMESH STEP 3: Execute TypeScript code that can call multiple MCP tools and process their results.\n\n<critical>\nTools are available as server objects with methods. Use serverName.methodName() format (e.g., await weatherServer.getForecast({ latitude: 36.5, longitude: -76.2 })). Do NOT use dire...
AI agents invoke execute-code to trigger actions in Codemesh. What it does depends on the arguments the agent supplies, and its effects often reach beyond the immediate call — builds kicked off, notifications sent, workflows started.
This tool allows executing arbitrary TypeScript code with access to call any method on any available MCP server. This is a code execution capability with unrestricted blast radius—an AI agent using this tool could invoke any side-effect-producing operation across all sibling tools (modify files, make queries, call external services).
From the tool's definition Tool description explicitly states 'Execute TypeScript code that can call multiple MCP tools and process their results.' The tool name is 'execute-code' and it provides runtime execution of arbitrary TypeScript with access to multiple MCP server methods.
Documented attack patterns abuse exactly the kind of access execute-code gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Codemesh, and nothing reaches the server without passing your rules. This is the rule we recommend for execute-code:
{
"version": "1",
"default": "deny",
"tools": {
"execute-code": {
"limits": [
{
"counter": "execute-code_rate",
"window": "minute",
"max": 10,
"scope": "grant"
}
]
}
}
}execute-code stays usable, but rate-capped — a runaway agent can't fire it dozens of times a minute. Everything else on the server is denied unless you say otherwise.
Free to start. No card required.
CODEMESH STEP 3: Execute TypeScript code that can call multiple MCP tools and process their results.\n\n<critical>\nTools are available as server objects with methods. Use serverName.methodName() format (e.g., await weatherServer.getForecast({ latitude: 36.5, longitude: -76.2 })). Do NOT use direct function calls or. It is categorised as a Execute tool in the Codemesh MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.
Register the Codemesh MCP server in PolicyLayer and add a rule for execute-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 Codemesh. Nothing to install.
execute-code is a Execute tool with high risk. Execute tools should be rate-limited and have argument validation enabled.
Yes. Add a rate_limit block to the execute-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 execute-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.
execute-code is provided by the Codemesh MCP server (kiliman/codemesh). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Start from Codemesh, 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.
13 Codemesh tools catalogued and risk-classified — across an index of 43,000+ MCP servers.