get_session_state¶

Science¶

A session is a credential, a profile, and a ledger. This tool returns the policy the caller is working under — view, subject, fence, budgets, review state — and a short list of calls already made. It does not return the payload of those calls. Cite the original tool_call_id, not this one.

The tool is exempt from total_calls so a caller can always ask how much budget remains. Time remaining and review-attempt counts are mechanical session facts, not a judgement of the event.

Contract¶

Generated from the live registry (ToolSpec). Field names, types and defaults come from the input and payload models; they are not hand-typed.


Group	`session`
Class	1
Tool version	`1.0.0`
Highest declared tier	F
Timeout	5 s
Budget key	`none`
`readOnlyHint`	true
`idempotentHint`	true

Input¶

Input schema has no fields.

Payload¶

Field	Type	Required	Notes
`profile`	string	yes
`view`	string ∈ {`blind`, `full`, `operator`}	yes
`subject`	object \| null	no	default `null`
`subject.event_id`	integer	yes
`expires_at_utc`	string	yes
`time_remaining_s`	integer	yes
`budgets`	object	yes
`fence`	object	yes
`fence.exclude_after_utc`	string \| null	no	default `null`
`fence.excluded_event_count`	integer	yes
`fence.exclude_decisions`	boolean	yes
`review`	object	yes
`review.status`	string ∈ {`not_submitted`, `rejected`, `accepted`}	yes
`review.attempts`	integer	yes
`review.max_attempts`	integer	yes
`calls`	array[object]	yes
`calls[].tool_call_id`	string	yes
`calls[].tool`	string	yes
`calls[].ok`	boolean	yes
`calls[].tier`	string ∈ {`F`, `E`, `D`}	yes
`calls[].catalogue_bearing`	boolean	yes
`calls[].hypothesis`	string \| null	no	default `null`
`calls[].summary`	string	yes	maxLength `120`

Example request¶

{}

Example payload¶

{
  "profile": "generic@1",
  "view": "full",
  "subject": {
    "event_id": 1001
  },
  "expires_at_utc": "2024-06-15T10:00:00.000Z",
  "time_remaining_s": 5410,
  "budgets": {
    "relocate": {
      "used": 1,
      "limit": 3
    },
    "total_calls": {
      "used": 17,
      "limit": 120
    }
  },
  "fence": {
    "exclude_after_utc": null,
    "excluded_event_count": 0,
    "exclude_decisions": false
  },
  "review": {
    "status": "not_submitted",
    "attempts": 0,
    "max_attempts": 3
  },
  "calls": [
    {
      "tool_call_id": "tc-0001",
      "tool": "get_event",
      "ok": true,
      "tier": "F",
      "catalogue_bearing": true,
      "hypothesis": null,
      "summary": "event 1001: 47.150N 24.500E 10.0 km ML 2.1, 9 phases"
    }
  ]
}

What this tool does not tell you¶

What this tool does not tell you: the contents of any result — only that a call happened and whether it succeeded. Cite the original tool_call_id, not this one.