Executes a tool — either a draft (by draftId) or a saved tool (by toolId), not both. Keep JSON payload data in payload and bind uploaded files separately in files. Set includeStepResults: true only when debugging wrong/empty output. Set returnFullConfig: true only when you need the full config.
AI agents invoke run_tool to trigger actions in Superglue MCP. 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 runs code/operations dynamically based on provided parameters. Although the actual impact depends on what the underlying tool does, the act of executing an arbitrary tool across 'apps, databases and APIs' creates substantial risk for side effects. Misuse could trigger unintended integrations, data modifications, or API calls.
From the tool's definition Tool description states it 'Executes a tool' and the context describes the server as enabling 'execution of integration pipelines automatically.' The tool triggers external operations (draft or saved tool execution) whose effects depend on arguments (draftId,…
Documented attack patterns abuse exactly the kind of access run_tool gives an agent:
PolicyLayer is an MCP gateway — it sits between your AI agents and Superglue MCP, and nothing reaches the server without passing your rules. This is the rule we recommend for run_tool:
{
"version": "1",
"default": "deny",
"tools": {
"run_tool": {
"limits": [
{
"counter": "run_tool_rate",
"window": "minute",
"max": 10,
"scope": "grant"
}
]
}
}
}run_tool 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.
Executes a tool — either a draft (by draftId) or a saved tool (by toolId), not both. Keep JSON payload data in payload and bind uploaded files separately in files. Set includeStepResults: true only when debugging wrong/empty output. Set returnFullConfig: true only when you need the full config. It is categorised as a Execute tool in the Superglue MCP MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.
Register the Superglue MCP server in PolicyLayer and add a rule for run_tool: 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 Superglue MCP. Nothing to install.
run_tool 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 run_tool 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 run_tool. 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.
run_tool is provided by the Superglue MCP server (superglue-ai/superglue). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.
Deterministic rules across all 21 Superglue MCP tools. Per-identity grants. Full audit log. Live in minutes. Nothing to install.
Free to start. No card required.
21 Superglue MCP tools catalogued and risk-classified — across an index of 42,500+ MCP servers.