Mind Chill
    Good Proof™by Mind Chill®
    HomeHow It WorksSectorsNewsMind Chill GuardiansPricing
    Book Sprint
    Mind Chill
    Good Proof™by Mind Chill®

    Contract-referenceable verification for high-impact AI actions. Scope-bound, expiry-aware, and human-final when it matters.

    Sales: [email protected]Security: [email protected]Support: [email protected]

    UK

    Mind Chill Nootropics Ltd

    09667911

    Singapore

    Mindchill Research Pte. Ltd.

    202544340Z

    A division of

    Mind Chill — Department of Human Defense

    Not a certification. Scope-limited verification. Acceptance depends on counterparty/programme requirements.

    Product

    • Good Proof Stamp
    • Stamp Spec
    • IDA Evidence Pack
    • How It Works
    • Verify API
    • Guardians
    • Pricing

    Solutions

    • Sectors
    • Specimens
    • Verify a Stamp
    • AI GOLD
    • Trust Metrics
    • RegTech
    • Security Automation

    Safeguards

    • Kill Switch
    • Agentic Security
    • Clause Pack
    • Coverage Reports
    • Portability & Data Rights

    Company

    • Book a Stamp Sprint
    • Advocate Partners
    • Partner Program
    • News
    • Leadership
    • Trust & Security
    • Official Domains

    © 2026 Good Proof by Mind Chill. All rights reserved.

    PrivacyTermsCookiesSecurityContactFAQStatusOfficial
    Transparency Layer

    AI GOLD Coverage Reports

    Programme-level operational metrics

    A programme-scoped public reporting layer for aggregate Good Proof throughput, published in phases. No sensitive payloads. No portals. Just measurable metrics.

    Fail-closedAggregate onlyProgramme-scoped
    Book a Stamp SprintView IDA Evidence Pack

    Why buyers use Coverage Reports

    Operational transparency without data exposure

    Audit/oversight-ready aggregate metrics

    Programme-scoped proof of fail-closed enforcement

    Definitions

    Programme

    A scoped verification deployment governed by specific rules, lanes, and eligibility criteria.

    Lane

    A category of high-impact actions sharing common verification requirements within a programme.

    Eligible Throughput

    Production Stamps in opted-in lanes that meet privacy thresholds and programme rules.

    Evidence Window

    The bounded time period during which a Stamp's verification status remains authoritative.

    Fail-Closed Event

    Verification unreachable returns NOT VERIFIED; the system fails safely rather than defaulting to trust.

    Validity Retention Rate

    Percentage of Stamps remaining VALID throughout their evidence windows without status degradation.

    Refresh Trigger

    A material-change event forcing NEEDS_REFRESH status, requiring re-verification before reliance.

    Withdrawal

    A stop-rely event where WITHDRAWN status is propagated, invalidating prior verification.

    Publishing phases

    1

    Specimen + methodology

    LIVE

    Redacted format example and methodology documentation available now.

    2

    Monthly publishing

    Monthly publishing begins once the programme meets [MIN_ELIGIBLE_THROUGHPUT_THRESHOLD] and privacy thresholds.

    3

    Lane sponsorship coverage + QA summaries

    Sponsored lane reports with sampling audit and QA summaries.

    Eligibility & Threshold Rules

    Contract-grade boundaries for what enters the transparency layer.

    Included

    • Production Stamps in defined lanes that meet [MIN_ELIGIBLE_THROUGHPUT_THRESHOLD]
    • Aggregate-only reporting fields (no case-level data)
    • Lanes explicitly opted into the transparency layer

    Excluded

    • Test traffic, staging environments, and synthetic data
    • Private programmes that have opted out
    • Lanes below [MIN_GROUP_SIZE] privacy threshold
    • Any content containing PII, PHI, or case details

    Opt-in/Opt-out: Private programmes may opt out entirely. Individual lanes may opt in or out per programme configuration.

    Methodology & Calculation Rules

    Auditor-ready metric definitions

    Reporting period

    Monthly calendar period (first day 00:00 UTC to last day 23:59 UTC).

    Stamp deduplication

    Each stamp_id counted once per reporting period; refreshes and status changes tracked separately.

    Validity retention rate

    Count of Stamps remaining VALID at window close ÷ total Stamps issued in period × 100.

    Fail-closed events

    Count of verification requests returning NOT_VERIFIED due to unreachable verification infrastructure.

    Methodology versioning

    Current version: v1.0. Changes published with revision notes; prior versions remain accessible.

    Change log policy: Methodology changes are published with revision notes; prior versions remain accessible for audit continuity.

    Published metrics

    Operational transparency tied to Good Proof primitives.

    Action counts

    Total Stamps issued within the reporting period across included lanes.

    Validity retention rate

    Percentage of Stamps remaining VALID through their evidence windows.

    Fail-closed events

    Verification unreachable → NOT VERIFIED. System failed safely.

    Refresh triggers fired

    Material-change events forcing NEEDS_REFRESH status.

    Withdrawals issued

    Stop-rely events where WITHDRAWN status was propagated.

    Guardian hours funded

    Human finality hours deployed for exceptions and dispute review.

    View Coverage Report Specimen

    Privacy & Aggregation Safeguards

    Security and legal protections

    • Aggregate only: no per-entity or case-level reporting
    • Small-cell suppression: categories below [MIN_GROUP_SIZE] are withheld or rolled up
    • Rounding/bucketing applied where necessary to prevent inference
    • Lanes failing privacy thresholds are withheld or aggregated into broader categories
    • No PII, PHI, case details, prompts, or logs published

    Integrity, Anti-Gaming & Assurance

    Underwriter-ready safeguards

    • Test, staging, and synthetic traffic excluded from all counts
    • Anomaly detection flags unusual issuance spikes for review
    • Sampling audits (when enabled) are programme-scoped with QA summaries
    • Reports are time-stamped artifacts traceable to programme records without exposing payloads

    Sponsor a Lane

    Fund verification capacity in specific impact categories

    Sponsorship funds capacity: Guardian hours, sampling audits, and lane QA for categories sponsors care about.

    • Fund Guardian hours and dispute resolution capacity
    • Support audit sampling and lane quality assurance
    • Recognition in published coverage reports

    Independence guarantee

    • Sponsors do not influence verification outcomes
    • Sponsors do not control sampling, thresholds, or report wording
    • Conflicts are checked; independence is documented programme-scoped
    • Sponsorship funds capacity (Guardian hours, sampling audits, QA), not outcomes
    Discuss Sponsorship

    Corrections & Dispute Process

    If errors are found in a published report, a corrected report is published with revision history and date. Prior versions remain accessible for audit trail continuity.

    Related resources

    AI GOLD for Duty of Care

    Duty-of-care evidence workflow

    IDA Evidence Pack

    Time-stamped artifacts you can forward

    Not a financial instrument. No promised returns. Metrics are operational transparency artefacts. Programme-scoped; eligibility rules apply.