For Trust & Safety Teams
Enforcement without evidence is liability.Evidence without scope is theatre.
Good Proof gates high-impact enforcement actions — suspensions, removals, de-platforming — with fail-closed verification and appeal-ready evidence at decision-time.
Scope-bounded enforcement • Appeal-ready evidence • Revocable authorization • Decision-time audit trail
Not a certification. Scope-limited verification.
Trust & Safety 60-Second Brief
Problem
Irreversible platform actions — suspensions, removals, de-platforming — without verifiable decision trails.
Control
Status Link verification before high-impact enforcement actions execute.
Output
Evidence Pack per enforcement decision for appeals, audits, and regulatory review.
Boundary
Scope-limited to enforcement actions. Does not replace content moderation models or policy design.
Is this a fit?
Best fit
- Platforms executing irreversible actions on user accounts, content, or access
- Teams that need appeal-ready evidence trails for regulatory or legal scrutiny
- Enforcement workflows where false positives carry material user harm
Not fit
- Content ranking or recommendation adjustments with no enforcement consequence
- Teams seeking content moderation model training or prompt tuning
Done means: enforcement actions block automatically when status ≠ VALID.
Enforcement Outcomes
Scoped enforcement gate
Suspension/removal only executes if verification returns VALID
Revocable authorization
Withdrawn status halts future enforcement reliance immediately
Appeal-ready evidence
Every enforcement decision produces an exportable Evidence Pack
Transparency signal
Affected users can verify the decision status via Status Link
How Enforcement Works
1
Enforcement action triggered (suspension, removal, restriction)
2
Gate verifies Status Link from authorized verifier
3
Policy evaluates status + scope + expiry against action class
4
Execute only if VALID — otherwise block and escalate to human review
5
Log Gate Decision + evidence fields for appeal and audit
Failure Semantics
Explicit condition → state → action mapping
| Condition | State | Action |
|---|---|---|
| Verifier timeout | NOT_VERIFIED | Block/Escalate |
| TLS failure | NOT_VERIFIED | Block/Escalate |
| Domain mismatch | NOT_VERIFIED | Block/Escalate |
| Malformed response / signature failure | NOT_VERIFIED | Block/Escalate |
| WITHDRAWN | WITHDRAWN | Block |
| NEEDS_REFRESH | NEEDS_REFRESH | Escalate/Review |
| Out-of-scope / expired | NOT_VERIFIED | Block/Escalate |
What a Stamp Proves
Proves (within scope)
- Enforcement action class + Gate Decision
- Verification timestamp at decision-time
- Signer/authority reference
- Scope boundaries + expiry window
- Validity state at moment of enforcement
- Evidence chain for appeal reconstruction
Does not prove
- Content moderation model accuracy
- Whether the policy itself is fair or proportionate
- User intent or behavioural prediction
- Raw content payloads triggering enforcement
- Regulatory compliance guarantee
In appeals: Status Link = current enforcement state. Evidence Pack = decision-time record for reconstruction.
Status Triggers
NEEDS_REFRESH
- Policy or enforcement rule change
- Model version update in moderation pipeline
- Jurisdiction or regulatory scope change
- Key rotation or credential refresh
- Incident opened pending internal review
WITHDRAWN
- Confirmed wrongful enforcement action
- Appeal upheld — original decision reversed
- Scope boundary breach in enforcement action
- Regulatory order requiring halt
- Critical vulnerability in enforcement pipeline
Evidence for Appeals & Audit
Fields per Gate Decision
evidence-fields.json
stamp_idaction_classdecisionstatusverified_atexpires_atverifier_domainsignerversionrequest_idreason_codefailure_mode
Designed for decision-time reconstruction and external review.
Fits Into Your Safety Program
Compensating control, not replacement
Content moderation and classification modelsHuman review queues and escalation pathsUser appeal and dispute workflowsRegulatory reporting and transparency reportsPlatform policy and community guidelines
What Trust & Safety Teams Ask
Does this replace our content moderation system?
No. Good Proof verifies enforcement actions, not content classifications. Your moderation models decide what violates policy — Good Proof gates the enforcement that follows.
Can affected users see the verification status?
Yes. Status Links can be surfaced in appeal workflows, giving affected users a verifiable reference for the enforcement decision.
What about false positives in automated enforcement?
Good Proof doesn't prevent false positives in classification. It ensures enforcement actions are scope-bounded, evidence-backed, and revocable — reducing the harm when false positives occur.
How does this help with regulatory scrutiny?
Evidence Packs provide decision-time reconstruction for regulators. Each enforcement action is traceable to a specific verification state, scope, and authority.
"Enforcement of high-impact [ACTION CLASS] is contractually conditioned on VALID status at decision time."
Verification path excludes raw user content and PII payloads by default.
"High-impact [ACTION CLASS] SHALL require a valid externally-verifiable Status Link (No Stamp → No Enforcement)."
30-Day Sprint
One enforcement lane live with fail-closed gating, Status Link verification, and Evidence Pack output.
1
Scope & criteria
Define 3 enforcement action classes, scope boundaries, and success criteria
2
Integrate gate
Wire Status Link verification into enforcement pipeline, logging, and appeal hooks
3
Failure + appeal testing
Test timeout/TLS/mismatch drills + wrongful enforcement reversal workflow
4
Go-live report
Metrics, false-positive impact data, appeal workflow validation, rollout plan
Risk Reduced vs Residual
Risk reduced
- Irreversible enforcement without decision-time evidence
- Appeal processes lacking verifiable audit trails
- Stale authorization persisting after policy or model change
- Regulatory exposure from untraceable enforcement decisions
Residual (out of scope)
- Content moderation accuracy or model bias
- Whether platform policies are fair or proportionate
- User behaviour prediction or intent analysis
- Blanket compliance or certification guarantees
Designed to ensure enforcement decisions are evidence-backed and reversible, not to evaluate policy fairness.
Assurance Governance (Optional by Lane)
Mind Chill Guardians
Human finality for high-impact enforcement exceptions
Conflict checks and reviewer rotation
Multi-review thresholds for permanent account actions
Minimal disclosure with auditable traceability
Ready to gate enforcement actions?
Definition of done: enforcement blocks on status ≠ VALID.
Every enforcement action without evidence is a future appeal you cannot defend.
Scope-limited verification. Not a certification.