Executes every step of a named chain (list names with find_chain / build_workflow_links) and returns ONE composite artifact whose execution_hash anchors all step outputs. compute:"server"/"auto" (default) runs each kernel-backed step server-side, threading step N's execution_hash into step N+1's ...
AI agents invoke run_chain to trigger actions in Ainumbers Mcp Apps. 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.
| Parameter | Type | Required | Description |
|---|---|---|---|
chain | string | Yes | Chain name, e.g. "agent-commerce-conformance". List names with find_chain or build_workflow_links. |
inputs | object | — | Map of step tool_id -> policy_parameters overrides. Omitted steps run with {} (kernels needing required fields are reported, not failed silently). |
compute | string | — | "auto"/"server" (default) runs kernel-backed steps server-side; "browser" returns a zero-egress delegation bundle to run client-side. |
Parameters from the server's own tool schema.
This tool executes arbitrary chains with server-side computation, threading outputs between steps. Although the server description claims 'read-only, no auth, zero PII', the tool itself performs programmatic execution of multi-step workflows with side effects that depend entirely on chain configuration and input parameters.
From the tool's definition Executes every step of a named chain...returns ONE composite artifact whose execution_hash anchors all step outputs. compute:"server"/"auto" (default) runs each kernel-backed step server-side, threading step N's execution_hash into step N+1's parent_hashes
Risk signalsBulk/mass operation — affects multiple targets
Attacks that exploit this kind of access
Executes every step of a named chain (list names with find_chain / build_workflow_links) and returns ONE composite artifact whose execution_hash anchors all step outputs. compute:"server"/"auto" (default) runs each kernel-backed step server-side, threading step N's execution_hash into step N+1's parent_hashes; compute:"browser" returns a zero-egress delegation bundle (composer URL + ordered deep-links) to run client-side instead — no data leaves the agent. Supply inputs as a map of step tool_id -> policy_parameters (field names per node manifest / build_chaingraph); a step whose kernel needs inputs you omit is reported per-step (status "input_required"), never failed silently. Steps that are browser-only (gpu:true or no registered kernel) are listed for browser delegation. Deterministic, zero PII, zero payload logging. Verify the result with verify_execution_hash. It is categorised as a Execute tool in the Ainumbers Mcp Apps MCP Server, which means it can trigger actions or run processes. Use rate limits and argument validation.
run_chain accepts 3 parameters: chain, inputs, compute. Required: chain. The full parameter table on this page comes from the server's own tool schema.
Register the Ainumbers Mcp Apps MCP server in PolicyLayer and add a rule for run_chain: 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 Ainumbers Mcp Apps. Nothing to install.
run_chain 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_chain 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_chain. 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_chain is provided by the Ainumbers Mcp Apps MCP server (postoaklabs/ainumbers-mcp-apps). 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 →