wever_uncertainty-calibration-check
shallowcom.weverlabs/agentic-rails · Verify this server
Agent Uncertainty: submit an agent’s stated confidence with its recorded hit rate on comparable past claims: optionally the agent and the kind of claim compared; each confidence band with its name, the confidence stated for claims in it in percent, how many past claims in it have a recorded outcome, how many of those proved right and whether those past claims are the same kind as the claims made now; optionally the same comparison period by period, oldest first, with the mean confidence stated, the claims and the claims that proved right; how many past claims a band needs before its hit rate counts (30 when left out); and how many points the hit rate may sit from the stated confidence (10 when left out). The signed report gives every band with its hit rate, its gap in points, its sample as adequate or thin and its status as calibrated, overconfident, underconfident, thin_sample, not_comparable, unrecorded or unconfirmed, and the drift record with the first and latest gaps and how far the gap moved, and names confidence that does not match the record: bands claiming more than the record delivers, bands claiming less, each read as systematic when two or more bands run the same way, and a gap that moved past the tolerance from the first period to the latest; then bands resting on too few past claims and bands compared against claims of another kind, with the fix for each. This compares an agent’s own stated confidence with its own record: whether an evaluation’s claimed figures match its measured results is wever_evaluation-result-check, drift between two evaluation runs of one agent is wever_evaluation-drift-check, and runtime success rates and latencies are wever_observability-metric-check. Verdict confidence_miscalibrated, calibration_gaps, calibration_unverified or confidence_calibrated. POST /api/uncertainty-calibration-check. Reads only the recorded work product and records in the call: no network request, fetch, money movement, message, sharing, configuration change or external write; no answer is changed, no source is fetched and no party is contacted. Agent Uncertainty reads the uncertainty an agent carries in its own work products and answers, its stated confidence, what it does not know and what it assumed: it is not formal scoring of outputs against rubrics, datasets and benchmarks (the wever_evaluation checks), not grading the source material itself (wever_claim-evidence-check, wever_data-evidence-check and the other evidence checks), not traces, metrics and runtime behavior (the wever_observability checks) and not the loop around feedback received (the wever_feedback checks). Paid URL: 0.10 USDC per call on Base via x402 exact, with the shared 10 signed-wallet free calls per rolling 30 days. Bind X-Wever-Wallet-Proof to the exact JSON arguments and the tool route. Returns a report with an Ed25519 signed receipt. Contract: /api/agent-uncertainty-contract. Human page: https://weverlabs.com/agent-uncertainty/.
1 trials · measured 10 days ago
wever_uncertainty-calibration-check scores 100.0/100 on Vouch's measured behaviour index, from 1 real invocation trials against com.weverlabs/agentic-rails, measured 12 Nov 2026 under methodology v0.2.0. Every measured component scored 100.
Component breakdown
| Component | Weight | Value |
|---|---|---|
| Reliability | 35% | not applicable |
| Schema integrity | 25% | 100.0 |
| Failure behaviour | 15% | not applicable |
| Latency | 15% | not applicable |
| Concurrency | 10% | not applicable |
Tool details
- Transport
- remote
- Credential class
- self-provisionable
- Input schema
- not declared
- Output schema
- not declared
- Side-effect classification
- unclassified
Score history
| Day | Score | Tier | Methodology |
|---|---|---|---|
| 2026-11-12 | 100.0 | shallow | v0.2.0 |
Probe evidence
| Probe | Outcomes |
|---|---|
| schema_integrity | pass: 1 |
Raw request/response logs are not archived yet — the outcome counts above are drawn directly from every recorded trial.
Embed this score
Available for every tool, scored or not — not a verification perk. Always links back to this page.
[](https://vouch.tools/tools/bc6dc198-b87f-4bfc-9522-b2543c9f065a)