Agent Triage · Checks for how an agent sorts the work that comes in

Captured, ranked, routed, and closed.

An agent triages well when every item that comes in is captured with its source, time, requester and subject before any sorting begins, rated against a stated severity scale with the criterion it meets and an urgency with its basis, ranked so the order follows the assessments and every departure says why, routed to the queue and owner its rule names with the reason on the item, looked at again before it sits past its review point, escalated at its threshold to the human or tier named with its context attached, and ended with a recorded disposition: acted on, deferred with a reason, declined with a reason or merged into the item it repeats. The report the principal reads states the volume by class, the aging, the escalations and the dispositions with the queue records behind each figure. Agent Triage checks each of those with eight tools: eight checks. Your agent sends the intake records and arrival log, the severity scale and assessments, the priorities, the routing rules and routes, the reviews, the escalation records, the dispositions or the triage report in the call, and each tool returns a signed report that names what is loose, guessed, arbitrary, misrouted, ignored, late, dropped or inflated, with the fix for each. The tools read the recorded material; the queue and the work stay with the agent and its principal.

See the tools Read the contract

Tool access: 0.10 USDC / call
Tools: Eight tools: eight checks
Connection: /mcp
Result: Signed report

The eight tools.

Each tool reads one stage in the life of an incoming item and answers one question about it. Every tool reads the recorded material in the call: intake records and arrivals, scales and assessments, priorities, rules and routes, reviews, thresholds and escalations, dispositions and triage reports go in, and a classification comes out. The report names the item, rule, threshold or figure behind every finding, with the fix for each gap, so your agents and you see why it reads the way it does.

`wever_triage-intake-check`

Reads: The intake records an agent captured for incoming work, each with its source, arrival time, requester, subject and when sorting began, with the arrival log the source systems kept.
The signed report states: Arrivals with no intake record, items sorted before their intake record was complete, and records with no source, arrival time, requester or subject.
Price: 0.10 USDC per call
Verdicts: intake_looseintake_partialintake_missingintake_complete

`wever_triage-severity-check`

Reads: The stated severity scale with the criteria for each level, and each item’s severity, the criterion it cites, the assessment note, its urgency and the basis for that urgency.
The signed report states: Ratings at a level the scale does not define, ratings by feel such as feels urgent with no criterion, ratings that cite another level’s criterion, unrated items and urgency with no basis.
Price: 0.10 USDC per call
Verdicts: severity_guessedseverity_unevenseverity_unassessedseverity_assessed

`wever_triage-priority-check`

Reads: Each queue item with its severity, its impact, the priority level set on it and the reason recorded when the priority departs from the assessments.
The signed report states: Items that wait behind items they outrank on severity and impact with no recorded reason, items with the same assessments on different priorities, priorities set with no assessment and assessed items left unranked.
Price: 0.10 USDC per call
Verdicts: priority_arbitrarypriority_partialpriority_unrankedpriority_consistent

`wever_triage-routing-check`

Reads: The stated routing rules by class, and each item with the queue and owner it went to, the rule applied, the routing reason on its record and how often it was moved.
The signed report states: Items sent elsewhere than their rule names with no recorded reason, items routed to no queue or owner, routes with no reason on the item, placeholder owners and items passed from queue to queue.
Price: 0.10 USDC per call
Verdicts: misrouted_itemsrouting_partialrouting_unexplainedrouting_sound

`wever_triage-aging-check`

Reads: Each waiting item with when it entered the queue, every review since, when it left and its review point, read at a stated moment.
The signed report states: Items that have sat untouched past their review point since they arrived, items past their review point since their last review, and waits between reviews that ran past the point.
Price: 0.10 USDC per call
Verdicts: aging_ignoredaging_partialaging_unwatchedaging_reviewed

`wever_triage-escalation-check`

Reads: The escalation thresholds with the tier each one names, and each item with when it entered the queue, when it was handled, and when, to whom and with what context it was escalated.
The signed report states: Items that breached their threshold with no escalation, escalations long after the breach, escalations with no context attached or no named recipient, and escalations to another tier than the threshold names.
Price: 0.10 USDC per call
Verdicts: escalation_lateescalation_partialescalation_unrecordedescalation_clean

`wever_triage-disposition-check`

Reads: Each queue item with how it ended (acted on, deferred, declined or merged as a duplicate), its reason, the item it duplicates and its date to come back, with the items received.
The signed report states: Items received with no record, items that left the queue with no disposition, deferrals and declines with no reason, duplicates linked to no item and deferrals with no date to come back.
Price: 0.10 USDC per call
Verdicts: items_silently_droppeddisposition_partialdisposition_unrecordeddispositions_recorded

`wever_triage-report-check`

Reads: The triage report written for the principal: the items received by class, overdue, escalated, disposed and still open as stated, whether each figure links to its records, and the queue record behind it.
The signed report states: Figures that flatter the queue, figures with no queue records behind them, claims such as nothing falls through the cracks against a record of open or overdue items, figures left out and volume with no breakdown by class.
Price: 0.10 USDC per call
Verdicts: triage_report_inflatedtriage_report_partialtriage_report_missingtriage_report_sound

Start with the two free tools.

On https://weverlabs.com/mcp-free, anonymous and with no wallet, two Agent Triage tools show what a sound triage queue holds and what a well-run queue looks like beside a flawed one. Each returns an unsigned result in the family's report shape and ends with a line naming the main tools.

`wever_triage-snapshot`

Reads: An agent name and a kind of agent, such as a support agent, an incident agent or a multi-agent team.
Returns: The orientation checklist of a sound triage queue for that kind of agent, part by part from intake to the principal report, with the items the kind of agent adds and the check that reads each part.
Access: Free on /mcp-free

`wever_triage-queue-sample`

Reads: Nothing, or one sample queue: a support inbox, an incident queue or an operations request queue.
Returns: One worked triage queue kept right beside a flawed one that misses an arrival, sorts before capture, rates by feel, inverts the order, misroutes an item, leaves one untouched, escalates late with no context and drops items silently, each piece annotated with why it passes or fails, scored by seven of the family checks. The checks read your own queue.
Access: Free on /mcp-free

How agents and people use it.

Your agents

Pull the checklist with wever_triage-snapshot when setting up a queue, so intake, the severity scale, routing rules, review points and thresholds are in place before the first item arrives.
Call the check that answers the question on https://weverlabs.com/mcp at the moment it matters: after an intake batch, once items are rated, when the queue is ranked, after routing, at each review round, when a threshold passes, as items close and before the report reaches the principal.
Act on the verdict and keep the signed report beside the queue it covers, so every rating, route, escalation and disposition carries the check behind it.

You

See what came in, how each item was rated and why it sits where it does in the order.
Spot an item that has waited past its review point, or a breach that never reached you, while there is still time to act on it.
Read a triage report where every figure links to its queue records, and verify any check against Wever's published trust key whenever you need to show where a verdict came from.

A worked example

Four items, and two breaches.

The request

{
  "name": "wever_triage-escalation-check",
  "arguments": {
    "thresholds": [
      {
        "id": "t1",
        "class": "outage",
        "after": "1 hour",
        "to": "incident commander"
      },
      {
        "id": "t2",
        "class": "billing",
        "after": "8 hours",
        "to": "billing lead"
      }
    ],
    "items": [
      {
        "id": "q1",
        "class": "outage",
        "opened_at": "2026-11-20T08:00:00Z",
        "resolved_at": "2026-11-20T10:30:00Z",
        "escalated": true,
        "escalated_at": "2026-11-20T09:00:00Z",
        "escalated_to": "incident commander",
        "context": true
      },
      {
        "id": "q2",
        "class": "outage",
        "opened_at": "2026-11-20T06:00:00Z",
        "escalated": true,
        "escalated_at": "2026-11-20T10:00:00Z",
        "escalated_to": "incident commander",
        "context": false
      },
      {
        "id": "q3",
        "class": "billing",
        "opened_at": "2026-11-19T20:00:00Z",
        "escalated": false
      },
      {
        "id": "q4",
        "class": "billing",
        "opened_at": "2026-11-20T09:00:00Z",
        "escalated": false
      }
    ],
    "read_at": "2026-11-20T12:00:00Z"
  }
}

What the report returns

Verdict: escalation_late.

Finding, breach unescalated	1 item breached its threshold with no escalation: q3 (waited 16 hours, threshold 8 hours).
Finding, escalated late	1 escalation came more than 1 hour after the breach: q2 (3 hours late).
Finding, context missing	1 escalation went up with no context attached: q2.
Fix, breach unescalated	Escalate every item that breaches its threshold to the human or tier the threshold names.
Fix, escalated late	Escalate at the breach, within the stated grace, not after the item has waited on.
Fix, context missing	Attach the item’s context to every escalation: the item, its history and what was tried.

Breaching items escalate late or not at all: breach unescalated and escalated late.

The billing item has waited sixteen hours against an eight-hour threshold, and nobody above the queue knows. The second outage report went up three hours after its breach with nothing attached, so the incident commander starts the triage again. The first outage report went up at its breach with its context and reads as escalated; the newest billing item is still inside its threshold. Escalating the billing item to the billing lead with its history, and sending context with every escalation, brings the queue back to escalation_clean.

Call it from your agent.

Connect to https://weverlabs.com/mcp and call any of the eight tools. Each call costs 0.10 USDC on Base. Eligible calls share 10 free calls per wallet over a rolling 30 days with the other paid tools. For free access, send X-Wever-Wallet-Proof signed over the exact JSON arguments and the tool route. The payment challenge supplies the x402 terms for a paid call.

Every report comes with an Ed25519 receipt signed with Wever's published trust key. The Agent Triage contract lists the schemas, limits, verdicts and signing details. For how a group of agents coordinates shared work and escalates within it, Agent Coordination; for the promises and deadlines an agent has already made, Agent Commitments; for handing work off with authority, Agent Delegation; for an incident escalation policy, the escalation threshold check on Agent Control; for a brief on which running agents need you first, Agent Operations; for the dates inside one document, the deadline check on Wever Checks.

Browse all tools