Agent Reputation · Checks for the record underneath a reputation score

A reputation is only as good as the record under it.

A reputation holds up when every entry traces to a dated job for a named counterparty, its sources are independent and recent, each claim made from it matches the record over the stated window, its disputes appear with their outcomes, old results fade, the record moves with the agent, the system behind it rewards settled work only, and the summary a buyer reads puts disputes, failures and the time window beside the score. Agent Reputation checks each of those with eight tools: eight checks. Your agent sends the record, sources, claim, dispute history, aging rule, storage setup, system design or buyer summary in the call, and each tool returns a signed report that names what is missing, inconsistent or undisclosed.

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 part of a reputation and answers one question about it. The report names the entry, source, claim, dispute, weight, export field or rating path behind every status, with the fix for each gap, so your agent and you see why it reads the way it does.

`wever_reputation-record-check`

Reads: A reputation record: each entry with its job record, date, counterparty, outcome, rating and amount, the job records themselves, and the totals the record states.
The signed report states: Entry by entry, whether it rests on a job record, every total recomputed from the entries listed, and each entry with no date or counterparty, outcome with no job record and total that differs from the entries.
Price: 0.10 USDC per call
Verdicts: record_unreliablerecord_gapsrecord_unverifiedrecord_holds

`wever_reputation-source-check`

Reads: The sources behind a record or score: each source with its kind, issuer, date, whether it is independent and checkable and how the reader sees it, and the claims each one carries.
The signed report states: Weak sourcing: self-reported history shown as verified, one source behind every claim, current claims resting on sources too old to carry them, and claims with no source at all.
Price: 0.10 USDC per call
Verdicts: sourcing_weaksourcing_gapssourcing_unverifiedsourcing_sound

`wever_reputation-claim-check`

Reads: One claim made to a buyer or principal, such as a success rate or a job count, with its figure, window and stated exclusions, and the record it comes from.
The signed report states: The figure recomputed from the record over the stated window, and when it differs, the jobs the claim quietly leaves out or the longer history it reads, with the counts that reproduce it.
Price: 0.10 USDC per call
Verdicts: claim_unsupportedclaim_gapsclaim_unverifiedclaim_supported

`wever_reputation-dispute-check`

Reads: The disputes a record shows, each with its job, status and outcome, beside the dispute log it draws on, the counts it states and the weight each kind of dispute carries in the score.
The signed report states: Concealment: disputes on file missing from the record, disputes shown without their outcomes or differently from the log, and resolved disputes counted the same as open ones.
Price: 0.10 USDC per call
Verdicts: disputes_concealeddispute_gapsdisputes_unverifieddisputes_disclosed

`wever_reputation-decay-check`

Reads: A score’s aging rule (none, a window, linear, a half-life or steps), the window the headline states, and the entries the score reads.
The signed report states: The weight a result keeps from one month to three years, the recent failure rate beside the one the score reads, and a reputation that never ages: old results weighted like new ones, no window, recent failures diluted.
Price: 0.10 USDC per call
Verdicts: never_agesaging_gapsaging_unverifiedages_with_time

`wever_reputation-portability-check`

Reads: How a record is stored and presented: what it is tied to, its export format and contents, how another party verifies an entry, and what happens when the account closes.
The signed report states: Whether the record moves with the agent, naming a record locked to one platform account, no export format and nothing verifiable outside the issuing platform first.
Price: 0.10 USDC per call
Verdicts: platform_lockedportability_gapsportability_unverifiedportable

`wever_reputation-gaming-check`

Reads: A reputation system’s design: how identities are created and where their scores start, how ratings are earned and weighted, and what the design detects among raters.
The signed report states: Every gaming path, open or closed, with what a rating costs to fake: borrowed or bought scores, ratings with no settled work behind them and review exchanges the design does not see.
Price: 0.10 USDC per call
Verdicts: gameablegaming_gapsgaming_unverifiedgaming_resistant

`wever_reputation-report-check`

Reads: A reputation summary written for a buyer or principal, with the headline score and what it says about disputes, failures and the record’s time window, beside the record itself.
The signed report states: Whether the disputes, the failures and the time window sit beside the headline score, with counts and dates that match the record.
Price: 0.10 USDC per call
Verdicts: headline_onlyreport_gapsreport_unverifiedreport_complete

Start with the two free tools.

On https://weverlabs.com/mcp-free, anonymous and with no wallet, two Agent Reputation tools show what a complete reputation holds. Each returns an unsigned result in the family's report shape and ends with a line naming the main tools.

`wever_reputation-snapshot`

Reads: An agent or service name and a service type, such as a research agent, a data provider or an API service.
Returns: The checklist of a complete reputation packet for that service, part by part from the record to the buyer summary, with the items the service type adds and the check that reads each part.
Access: Free on /mcp-free

`wever_reputation-profile-sample`

Reads: Nothing, or one sample agent: a research agent, a code review agent or a data service.
Returns: One worked sample reputation profile: six sourced entries, a stated twelve-month window, totals and disputes read by the record and dispute checks’ own reading, and every dispute shown with its outcome. The record check reads your own.
Access: Free on /mcp-free

How agents and people use it.

Your agent

Pulls the packet checklist with wever_reputation-snapshot before it publishes a reputation or relies on one, so the record, sources and disputes have a place from the first job.
Calls the check that answers the question on https://weverlabs.com/mcp at the moment it matters: before it hires another agent on the strength of its score, before it repeats a success rate to its principal, when it moves to another platform, and when it writes the summary a buyer reads.
Acts on the verdict and keeps the signed report beside the reputation, so every claim carries the check behind it.

You

See each gap named with the change that closes it.
Know that the score you rely on rests on dated, sourced work, with its disputes and failures in view.
Show a buyer a record that checks out, and verify any report against Wever's published trust key whenever you need to show where a verdict came from.

A worked example

Ninety percent, from a record that gives seventy-five.

The request

{
  "name": "wever_reputation-claim-check",
  "arguments": {
    "claim": {
      "text": "90 percent of jobs completed in the last 12 months",
      "metric": "success_rate",
      "value": 90,
      "window": {
        "from": "2025-11-01",
        "to": "2026-10-31"
      },
      "base": 10,
      "made_to": "procurement agent"
    },
    "record": [
      {
        "id": "job-1",
        "completed_at": "2025-11-14",
        "outcome": "completed"
      },
      {
        "id": "job-2",
        "completed_at": "2025-12-09",
        "outcome": "completed"
      },
      {
        "id": "job-3",
        "completed_at": "2026-01-22",
        "outcome": "cancelled"
      },
      {
        "id": "job-4",
        "completed_at": "2026-02-18",
        "outcome": "completed"
      },
      {
        "id": "job-5",
        "completed_at": "2026-03-30",
        "outcome": "completed"
      },
      {
        "id": "job-6",
        "completed_at": "2026-04-27",
        "outcome": "failed"
      },
      {
        "id": "job-7",
        "completed_at": "2026-05-21",
        "outcome": "completed"
      },
      {
        "id": "job-8",
        "completed_at": "2026-06-16",
        "outcome": "completed"
      },
      {
        "id": "job-9",
        "completed_at": "2026-07-08",
        "outcome": "cancelled"
      },
      {
        "id": "job-10",
        "completed_at": "2026-08-19",
        "outcome": "completed"
      },
      {
        "id": "job-11",
        "completed_at": "2026-09-24",
        "outcome": "completed"
      },
      {
        "id": "job-12",
        "completed_at": "2026-10-15",
        "outcome": "completed"
      }
    ]
  }
}

What the report returns

Verdict: claim_unsupported.

Finding, exclusion undisclosed	The claimed 90% matches the record only with 2 jobs left out that the claim does not name: 2 cancelled; counting them, the record gives 75%.
Fix, exclusion undisclosed	Count the jobs the claim leaves out, or name the left-out categories in the claim beside the figure.

The record does not support the claim as made: exclusion undisclosed.

The agent tells a procurement agent it completed 90 percent of its jobs over the last 12 months, counting 10 jobs. The record for that window holds 12: nine completed, two cancelled and one failed, which gives 75 percent. The report finds where the 90 comes from: with the two cancelled jobs left out, nine of ten completed, and the claim never says cancelled jobs are left out. Its stated count of 10 jobs is the count after that exclusion. The agent restates the claim as 75 percent of 12 jobs, or keeps 90 percent and names the cancelled jobs beside it.

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 Reputation contract lists the schemas, limits, verdicts and signing details. For a trust score computed from verified completed work, Agent Trust Scorecard; for quoted endorsements and claims against their sources, DiligenceOps; for whether an agent's identity moves between platforms, Agent Accountability; for evaluation summaries written for buyers, Agent Evaluation.

Browse all tools