File a capability request against the vaultpilot-mcp GitHub repository when the user asks for something this server cannot do (e.g. an unsupported protocol, chain, token, or missing tool). USE ONLY AFTER confirming no existing tool can accomplish the task. By default this returns a pre-filled Git...
AI agents call request_capability to permanently remove resources in VaultPilot MCP — typically in cleanup and lifecycle workflows. It does its job in a single call, and there is no undo.
| Parameter | Type | Required | Description |
|---|---|---|---|
context | object | — | |
summary | string | Yes | One-line title of the missing capability (used as the GitHub issue title). E.g. 'Support Aerodrome LP positions on Base' or 'Add Pendle PT/YT position reader'. |
category | string | — | Rough bucket to help triage. |
agentName | string | — | MCP client identifier (e.g. 'Claude Code', 'Cursor'). Helps triage. |
description | string | Yes | What the user asked for, what the agent tried, what's missing, and why the existing tools don't cover it. Include protocol name, chain, contract addresses, and |
Parameters from the server's own tool schema.
An AI agent that decides to call request_capability doesn't hesitate, doesn't double-check, and doesn't stop at one. Whatever it removes from VaultPilot MCP is gone — there is no undo for destructive operations.
Attacks that exploit this kind of access
File a capability request against the vaultpilot-mcp GitHub repository when the user asks for something this server cannot do (e.g. an unsupported protocol, chain, token, or missing tool). USE ONLY AFTER confirming no existing tool can accomplish the task. By default this returns a pre-filled GitHub issue URL — NO data is transmitted; the user must click through to submit. If the operator has configured VAULTPILOT_FEEDBACK_ENDPOINT, it posts directly to that proxy instead. Rate-limited per install (30s between calls, 3/hour, 10/day, 7-day dedupe on identical summaries). Write clear, actionable summaries — this lands in a real issue tracker read by humans. It is categorised as a Destructive tool in the VaultPilot MCP MCP Server, which means it can permanently delete or destroy data. Block by default and require explicit approval.
request_capability accepts 5 parameters: context, summary, category, agentName, description. Required: summary, description. The full parameter table on this page comes from the server's own tool schema.
Register the VaultPilot MCP server in PolicyLayer and add a rule for request_capability: 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 VaultPilot MCP. Nothing to install.
request_capability is a Destructive tool with critical risk. Critical-risk tools should be blocked by default and only enabled with explicit human approval.
Yes. Add a rate_limit block to the request_capability 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 request_capability. 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.
request_capability is provided by the VaultPilot MCP server (vaultpilot-mcp). PolicyLayer sits as a proxy in front of this server to enforce policies before tool calls reach the server.