get_execution_context
Discover this account’s revalidated active roots, adapters, and asset selections.
| Field | Meaning |
|---|---|
| Side effects | None. This is a read-only execution call. |
| Next step | Build one complete supported request and call propose_execution with the selected root. |
Arguments
| Argument | Type | Required |
|---|---|---|
cursor | string | no |
page_size | integer | no |
Exact bounds, accepted values, and nested fields are in Full JSON Schema below.
Example
Select one exact root from active_pakts. Continue through next_cursor if necessary. This read does not resolve wallet readiness or start authorization.
With an authenticated MCP client:
await client.callTool({
name: "get_execution_context",
arguments: { page_size: 10 },
});Returns
Important fields in the decoded result; this is not a complete response schema.
| Field | Meaning |
|---|---|
active_pakts[] | Revalidated roots and their execution_adapter and policy_program_id. |
active_pakts[].policy_program_generation | For Hyperliquid: current, account_leverage_v6, mixed_margin_mode_v5, single_margin_mode_v4, or cross_only_v3; omitted for other adapters. |
active_pakts[].asset_selection | For Hyperliquid: "any" or an exact only list with chain, asset index, and oracle. |
active_pakts[].policy_rules[] | Committed limits and refusal conditions (rule, refused_when, trust basis), plus display-only headline (for example ≤ $50 per trade, or null), rule_core without the reduce-only close exemption, close_exempt, and boundaries: the rule’s rows under can, never, or stops_when, each with a title and description. |
wallet_readiness / execution_readiness | checked_on_proposal; active does not mean ready to sign. |
account_id / environment / next_cursor | Authenticated scope and pagination cursor. |
For error handling and retry rules, see Statuses and Errors.
Read one bounded page of this authenticated account's active, fully revalidated pakt roots, artifact-derived adapters, proposal-relevant policy_rules, and asset selection where applicable. policy_rules is the canonical SSZ-derived readback of numeric limits and refusal conditions; inspect it before constructing a proposal. Each policy_rules entry also carries display-only reader fields from the same readback: headline is the committed limit as a short value such as ≤ $50 per trade, or null when the rule has none; rule_core is rule without the reduce-only close exemption; close_exempt states whether authenticated position-bounded reduce-only closes are exempt from the rule; boundaries lists the rule's rows among the agent's boundaries, each with a section (can, never, or stops_when), a title, and a description without the close exemption. Services additionally expose services_policy: version, valid_until as a decimal string, service_permissions and expiry_config arrays, and for a paying services Pakt service_payment: payments_wallet, tokens and payees with max_per_payment_atomic, and unlisted_payee_max_per_payment_atomic, atomic amounts as decimal strings. Every entry includes its canonical scope; USD micro-unit amounts are decimal strings. This is canonical artifact configuration, not mutable spend history or an execution allowance. A Hyperliquid asset_selection is either any or an exact only-list of native indices on the named network, never Pakt-specific ids. It does not claim a planner or strategy capability. AccountId and environment are server-derived. Wallet and operation-authorization readiness are checked only when propose_execution enters the configured hosted authorizer. This call is read-only and never creates an account, provisions a wallet, proves, signs, or submits.
From tools/list, MCP protocol revision 2026-07-28.
{
"additionalProperties": false,
"properties": {
"cursor": {
"description": "Exclusive pakt-root cursor returned as next_cursor by the preceding page.",
"maxLength": 64,
"minLength": 64,
"pattern": "^[0-9a-f]{64}$",
"type": "string"
},
"page_size": {
"default": 10,
"maximum": 10,
"minimum": 1,
"type": "integer"
}
},
"type": "object"
}