Trust policy
Corrections remain visible and attributable
GridNinja retains a permanent record when a substantive public claim is corrected, expires, or is retracted. The record identifies the affected wording and URL, prior state, reason, date, evidence change, owner, and replacement language when applicable. Corrections do not silently erase prior claims, and retracted values are removed from snippets, structured data, and active promotional surfaces.
Definition
What it is
A correction changes inaccurate or incomplete wording while preserving provenance; a retraction withdraws a claim that can no longer be supported; expiry pauses use pending review.
Category boundary
What it is not
It is not a general page changelog for typography or layout updates, nor a mechanism to hide unsupported claims after they have circulated.
Operational significance
Why operators care
A durable history lets operators and partners understand what changed and whether prior decisions or downloaded artifacts require re-evaluation.
Mechanism
How the decision path works
- Step 01
Freeze the affected claim record and identify every approved surface.
- Step 02
Publish the prior wording, reason, evidence change, and effective date.
- Step 03
Remove or replace active wording and structured facts.
- Step 04
Link superseded artifacts and notify owners of dependent material.
Textual diagram equivalent
Decision flow
Issue discovered → affected claim frozen → correction or retraction record → surfaces and artifacts updated → dependent reviews notified → permanent linked history.
Correction record example
Visibly scoped example
If a benchmark fixture is found to share a faulty assumption with its evaluator, the result is retracted, the artifact remains linked as superseded, and replacement evidence receives a new version.
Failure and no-proof cases
- A value disappears without a public record after external distribution.
- Structured data retains a fact removed from visible copy.
- A superseded artifact is overwritten at the same versioned URL.
Limitations
- No public correction entries exist at launch; an empty history must not imply perfect accuracy.
- Notification reach may be limited for independently copied material.
- Named editorial ownership is required before this policy becomes indexable.
Evidence and sources
Inspect the basis, not just the answer
Local evidence objects show the GridNinja proof contract. Primary sources provide external standards and risk-management context; they do not validate GridNinja performance.
Related operator resources
Continue the proof path
- Claims must never outrun evidenceHow GridNinja scopes public claims by environment, maturity, evidence, approved surface, caveat, and review expiry.
- A scoped, source-backed comparison policyHow GridNinja scopes public comparisons to reviewed primary materials, defined categories, stable citations, and explicit limitations.
- Evidence for safe, usable, auditable capacityPublication-gated methods, synthetic traces, ledgers, specifications, and proof artifacts for evaluating virtual capacity claims.
Proof before autonomy
Test the evidence boundary against your capacity question
Start with a Capacity Audit or read-only Shadow Mode discussion. GridNinja does not require control authority to identify where a capacity claim remains unproven.