# CHQ Criteria Governance Standard

- **Artifact ID:** CHQ-CGS-2026-001
- **Version:** v1.0
- **Status:** IN FORCE (effective 2026-09-22, Issue No. 39); FROZEN
- **Public record:** https://record.cybersecurityhq.com/doctrine/chq-cgs-2026-001
- **Machine-record SHA-256:** `0a2f929d9a62788185cf2fd5cf479c8a45117b5a29ad92883f2edda820a2d9d4`

## Complete structured record

```json
{
  "id": "CHQ-CGS-2026-001",
  "title": "CHQ Criteria Governance Standard",
  "version": "v1.0",
  "issued": "2026-09-15",
  "status": "IN FORCE (effective 2026-09-22, Issue No. 39); FROZEN",
  "doctrine_class": "GOVERNANCE_STANDARD",
  "doctrine_scope": "CONDITIONS",
  "governed_artifacts": [
    "Structural Conditions",
    "Rating Actions",
    "Structural Condition Board"
  ],
  "related_doctrine": [
    "CHQ-D-2026-PIG",
    "CHQ-D-2026-ESG"
  ],
  "purpose": "This Standard states how CHQ is permitted to create, validate, apply, change, invalidate, and retire the criteria that move ratings, and how evidence is graded before it may do so.",
  "canonical_text": "**CYBERSECURITYHQ // RECORD**\n\n# CHQ Criteria Governance Standard\n\n**CGS v1.0 · Declared 15 September 2026 · Effective 22 September 2026 (Issue No. 39) · Status: FROZEN**\n\n*The Structural Conditions Registry states what each condition means. This Standard states how CHQ is permitted to create, validate, apply, change, invalidate, and retire the criteria that move ratings, and how evidence is graded before it may do so. Every provision carries an identifier. From Issue No. 39, rating actions cite the provisions they were taken under. The Standard is frozen at v1.0; it changes only under Section 11, and never to fit a case already observed.*\n\n---\n\n## CGS-0 · Governing principle\n\n**CGS-0.1** No observation may determine both the rule by which it produces a rating action and that rating action. Evidence may reveal that a rule is defective and cause a replacement rule to be written; the replacement rule does not use that observation to produce an action in the review that created it.\n\n**CGS-0.2** Procedural consistency is necessary and not sufficient. A criterion applied identically to every case demonstrates that the rule ran; it does not demonstrate that the rule measures the condition it names.\n\n**CGS-0.3** Every rating is reconstructable as a chain: evidence, admissibility grade, criterion version, state transition, rating, publication. A rating that cannot be reconstructed through that chain is not defended; it is placed under review.\n\n---\n\n## CGS-1 · Criterion validation before activation\n\n**CGS-1.1** No criterion may affect a rating until it has passed the validation protocol below, with answers recorded in the registry.\n\n**CGS-1.2** Construct: what the condition claims exists, in one sentence.\n\n**CGS-1.3** Observable: the measurable event the criterion counts.\n\n**CGS-1.4** Link: why the observable measures the construct, stated as a mechanism, not an association.\n\n**CGS-1.5** False-positive test: the circumstances under which the observable occurs while the condition is absent. If a common circumstance satisfies the observable without the construct, the criterion fails.\n\n**CGS-1.6** False-negative test: whether the condition can exist without producing the observable, and whether that gap is acceptable.\n\n**CGS-1.7** Boundary test: cases immediately above and below the threshold must classify sensibly.\n\n**CGS-1.8** Historical test: the criterion is applied to known cases without alteration before activation, per CGS-7.\n\n**CGS-1.9** A criterion that fails any test is not activated. The record of the failed validation is retained.\n\n---\n\n## CGS-2 · Evidence hierarchy\n\n**CGS-2.1** Grade A, authoritative primary: government catalog data files and alerts, CVE records, securities filings, vendor security advisories, court and regulatory records.\n\n**CGS-2.2** Grade B, direct technical: vendor research with telemetry, incident-response reports with direct observation, original research with reproducible evidence, an operator's own incident postmortem.\n\n**CGS-2.3** Grade C, high-quality secondary: established news organizations and security publications reporting on Grade A or B material.\n\n**CGS-2.4** Grade D, unverified or indirect: aggregation of other sources, social posts, anonymous or uncorroborated claims, recycled content.\n\n**CGS-2.5** A rating-changing event requires Grade A or B evidence for each proposition the criterion depends on.\n\n**CGS-2.5a** Required propositions may be established by different admissible sources. No single source is required to establish the complete criterion unless the condition-specific criterion says otherwise.\n\n**CGS-2.6** Grade C may open a candidate observation and may enrich a counted instance; it cannot independently trigger reclassification.\n\n**CGS-2.7** Grade D may initiate investigation and nothing else. It does not enter the counted set, the ledger, or the chronology record.\n\n**CGS-2.8** A research pass, however deep, is a sweep and not a source; its findings are graded by the sources it cites, and its counts are verified against Grade A before use.\n\n---\n\n## CGS-3 · Admissibility\n\n**CGS-3.1** Evidence is admissible only for propositions the source directly supports. Source authority does not transfer to claims the source does not make.\n\n**CGS-3.2** Exploitation, actor identity, campaign linkage, motivation, and impact are separate propositions, each graded independently. Attribution does not inherit the grade of the exploitation it attaches to.\n\n**CGS-3.3** Where a criterion depends on event ordering, the registry records each date independently with its source: T_event, Evidence(T_event), for every event the criterion names. Inferred chronology cannot substitute for documented chronology unless the criterion explicitly permits inference. Number ranges, advisory serials, and adjacent evidence do not adjudicate a date.\n\n**CGS-3.4** Identifiers are matched in full form against admission fields only. A mention in a note is not an admission. A flag may be raised on Grade C or D; a miss may be declared only on Grade A dating.\n\n**CGS-3.5** An observation's founding instances must exist as signal rows before the observation opens.\n\n**CGS-3.6** Multi-identifier signals carry every identifier in the admission key.\n\n---\n\n## CGS-4 · Criterion change\n\n**CGS-4.1** Clarification: wording changes that alter no case's classification. Recorded; no version increment; no effective-date delay.\n\n**CGS-4.2** Refinement: a change to measurement that preserves the construct. Version increments; declared with an effective date no earlier than the next scheduled review.\n\n**CGS-4.3** Replacement: the prior criterion is found not to measure the construct. Version increments; the prior version is invalidated under CGS-8.\n\n**CGS-4.4** A revised criterion may not produce a rating action from the observation that caused its revision during the review in which the revision is declared. After the revised criterion becomes effective, the observation may be classified retrospectively for record completeness, but it does not independently satisfy a rating trigger unless the criterion expressly defines historical accrual.\n\n**CGS-4.5** Every substantive change records: version, declaration date, effective date, change type, reason, expected directional effect on qualifying observations, affected conditions, and retrospective test results per CGS-7.\n\n**CGS-4.6** No criterion is refined or replaced while an evaluation window it governs is open, except following invalidation under CGS-8. Invalidation takes effect immediately; a successor criterion follows CGS-4.5 and does not take effect before its declared effective date. Windows close on the scheduled review date.\n\n**CGS-4.7** A criterion for a state not yet reached, including retirement, is declared before the state is first reached, never at need.\n\n---\n\n## CGS-5 · Exceptional review\n\n**CGS-5.1** Review outside the schedule is permitted only on: a criterion found invalid; an authoritative source materially correcting underlying evidence; a duplicate observation discovered; a failure of evidence provenance; a state-machine ambiguity; or previously unavailable Grade A or B evidence that materially changes a case.\n\n**CGS-5.2** Exceptional review does not permit discretionary reclassification because evidence is interesting, surprising, or contrary to the board's thesis. The trigger is named in the record before the review begins.\n\n**CGS-5.3** Proximity to a trigger raises the evidence bar for the candidate that would meet it; it never lowers it.\n\n---\n\n## CGS-6 · Conflict and independence\n\n**CGS-6.1** When admissible sources conflict on a material fact, direct evidence of the proposition controls over indirect evidence regardless of general source grade. Where both sources contain direct evidence, the fact remains contested unless the discrepancy can be reconciled from the underlying record. Source grade alone does not resolve conflicting direct evidence: grade governs admissibility; evidentiary proximity governs conflict.\n\n**CGS-6.2** Multiple reports derived from one incident, vendor statement, telemetry set, or research document count as one evidentiary instance unless independently corroborated.\n\n**CGS-6.3** The counting unit for each condition is stated in its criterion. For the vendor risk-signal ledger the unit is the vendor assessment event. For exploitation conditions the unit is the sub-class instance. For containment and adversarial operations the unit is the distinct operation.\n\n**CGS-6.4** A sub-class is ratified on a definition with clean exclusions writable in one sentence, never on a shared noun, and never on count alone.\n\n**CGS-6.5** The taxonomy declares its gaps. An instance that fits no declared sub-class is recorded unattributed rather than forced.\n\n---\n\n## CGS-7 · Retrospective testing\n\n**CGS-7.1** Before activation, a criterion is tested against a set of historical observations covering known positives, known negatives, boundary cases, ambiguous cases, and adversarial cases.\n\n**CGS-7.1a** The test set contains all reasonably available qualifying historical cases when fewer than ten are known. Where ten or more cases are available, the test set contains at least ten and includes every category named in CGS-7.1 where such a case exists.\n\n**CGS-7.2** Results are published in three categories: QUALIFYING, NON-QUALIFYING, INDETERMINATE. Where the evidence cannot establish what the criterion needs, the case is indeterminate. Cases are not forced into yes or no.\n\n**CGS-7.3** The test set and results are recorded with the criterion version and are not revised after activation.\n\n---\n\n## CGS-8 · Invalidation and review states\n\n**CGS-8.1** A criterion is invalidated when its observable is found not to measure the construct it was designed to assess, or when a material defect makes its classifications unreliable.\n\n**CGS-8.2** An invalidated criterion produces no further rating actions.\n\n**CGS-8.3** A rating action whose sole qualifying basis was an invalidated criterion is VOIDED and the condition is placed UNDER REVIEW. The action is not reversed: reversal would decide the correct rating after the fact. The prior valid rating stands as the displayed state. The voided action, its date, and its criterion version remain visible in the record.\n\n**CGS-8.4** Every other adjudication made under the invalidated criterion is re-run under the valid successor at the same review.\n\n**CGS-8.5** Reclassification of a condition under review occurs only under a valid criterion at the next scheduled review or an exceptional review authorized under CGS-5.\n\n**CGS-8.6** The board displays a condition under review with: current displayed rating (the prior valid state), the voided action and its version, and the date and criterion version of the next adjudication.\n\n**CGS-8.7** A trigger met while a condition is at the top of the scale does not create a tier or a modifier. It is recorded, counted, and opens a Position-candidacy review under the issuance gate.\n\n**CGS-8.8** On de-escalation, prior adverse-event counts become historical context; re-escalation requires new events after the de-escalation.\n\n---\n\n## CGS-9 · Intake and reconciliation\n\n**CGS-9.1** Discovery runs against the authoritative data file, not against search or mirrors. Search and mirrors enrich; they do not discover.\n\n**CGS-9.2** Reconciliation runs source-to-registry across the full covered era, monthly, by full-form identifier.\n\n**CGS-9.3** The completeness-era boundary is a declared reconstruction. It is not moved to reclassify any absence.\n\n**CGS-9.4** Missed runs are declared gaps, never backdated. A gap obligates a source-file sweep of its window.\n\n**CGS-9.5** Every miss is classified, and every miss class carries a rule. The taxonomy is published.\n\n---\n\n## CGS-10 · Publication\n\n**CGS-10.1** Each issue cites the applied criteria by version and the provisions of this Standard under which each rating action was taken.\n\n**CGS-10.2** Board statistics carry rows for rating changes, scope refinements, watch changes, published triggers newly met with rating unchanged, rating actions voided, methodology changes, and corrections to prior issues. Rows are added when the machinery produces an event no row can hold; rows are never removed.\n\n**CGS-10.3** A correction to a prior issue is published at the volume of the original claim. An intake miss is disclosed with its class, its latency, and the rule that closes it.\n\n**CGS-10.4** Certainty is stated at the grade of the evidence. Reported is not confirmed; a label is not a chronology; a candidate is not a count.\n\n---\n\n## CGS-11 · Governance of the Standard\n\n**CGS-11.1** Changes to this Standard increment its version.\n\n**CGS-11.2** A CGS amendment records the provision changed, prior text, replacement text, reason, declaration date, effective date, and expected procedural effect.\n\n**CGS-11.3** No CGS amendment applies retroactively to alter the validity of an action taken under the Standard in force at the time.\n\n**CGS-11.4** An amendment arising from a live case cannot govern that case unless necessary to resolve a procedural state for which the Standard contains no rule. In that circumstance the amendment may govern procedure but cannot determine the substantive rating outcome.\n\n**CGS-11.5** Superseded versions remain permanently available.\n\n---\n\n## Application on declaration\n\n**SC-2026-006** is displayed under CGS-8.6 from this Standard's declaration: current rating STRENGTHENING (prior valid state); voided action CONFIRMED, 12 September 2026, criterion v1.0; next adjudication 24 September 2026 under criterion v2.0, whose CGS-1 validation and CGS-7 retrospective test are recorded in the registry before that date. Per CGS-4.4, the Fortinet observation that exposed v1.0 produces no rating action under v2.0 in the review that declared it; after v2.0 takes effect it is classified retrospectively for record completeness and does not independently satisfy a trigger.\n\n*End of Standard. CGS v1.0 is frozen. Provision identifiers are permanent.*",
  "canonical_hash": "792ad7c72644cc01a9e4cc7fc2a2e7220440912f37d44f0e5520883d4c734738",
  "sections": [
    {
      "heading": "Preamble",
      "content": "**CYBERSECURITYHQ // RECORD**\n\n# CHQ Criteria Governance Standard\n\n**CGS v1.0 · Declared 15 September 2026 · Effective 22 September 2026 (Issue No. 39) · Status: FROZEN**\n\n*The Structural Conditions Registry states what each condition means. This Standard states how CHQ is permitted to create, validate, apply, change, invalidate, and retire the criteria that move ratings, and how evidence is graded before it may do so. Every provision carries an identifier. From Issue No. 39, rating actions cite the provisions they were taken under. The Standard is frozen at v1.0; it changes only under Section 11, and never to fit a case already observed.*\n\n---"
    },
    {
      "heading": "CGS-0 · Governing principle",
      "content": "**CGS-0.1** No observation may determine both the rule by which it produces a rating action and that rating action. Evidence may reveal that a rule is defective and cause a replacement rule to be written; the replacement rule does not use that observation to produce an action in the review that created it.\n\n**CGS-0.2** Procedural consistency is necessary and not sufficient. A criterion applied identically to every case demonstrates that the rule ran; it does not demonstrate that the rule measures the condition it names.\n\n**CGS-0.3** Every rating is reconstructable as a chain: evidence, admissibility grade, criterion version, state transition, rating, publication. A rating that cannot be reconstructed through that chain is not defended; it is placed under review.\n\n---"
    },
    {
      "heading": "CGS-1 · Criterion validation before activation",
      "content": "**CGS-1.1** No criterion may affect a rating until it has passed the validation protocol below, with answers recorded in the registry.\n\n**CGS-1.2** Construct: what the condition claims exists, in one sentence.\n\n**CGS-1.3** Observable: the measurable event the criterion counts.\n\n**CGS-1.4** Link: why the observable measures the construct, stated as a mechanism, not an association.\n\n**CGS-1.5** False-positive test: the circumstances under which the observable occurs while the condition is absent. If a common circumstance satisfies the observable without the construct, the criterion fails.\n\n**CGS-1.6** False-negative test: whether the condition can exist without producing the observable, and whether that gap is acceptable.\n\n**CGS-1.7** Boundary test: cases immediately above and below the threshold must classify sensibly.\n\n**CGS-1.8** Historical test: the criterion is applied to known cases without alteration before activation, per CGS-7.\n\n**CGS-1.9** A criterion that fails any test is not activated. The record of the failed validation is retained.\n\n---"
    },
    {
      "heading": "CGS-2 · Evidence hierarchy",
      "content": "**CGS-2.1** Grade A, authoritative primary: government catalog data files and alerts, CVE records, securities filings, vendor security advisories, court and regulatory records.\n\n**CGS-2.2** Grade B, direct technical: vendor research with telemetry, incident-response reports with direct observation, original research with reproducible evidence, an operator's own incident postmortem.\n\n**CGS-2.3** Grade C, high-quality secondary: established news organizations and security publications reporting on Grade A or B material.\n\n**CGS-2.4** Grade D, unverified or indirect: aggregation of other sources, social posts, anonymous or uncorroborated claims, recycled content.\n\n**CGS-2.5** A rating-changing event requires Grade A or B evidence for each proposition the criterion depends on.\n\n**CGS-2.5a** Required propositions may be established by different admissible sources. No single source is required to establish the complete criterion unless the condition-specific criterion says otherwise.\n\n**CGS-2.6** Grade C may open a candidate observation and may enrich a counted instance; it cannot independently trigger reclassification.\n\n**CGS-2.7** Grade D may initiate investigation and nothing else. It does not enter the counted set, the ledger, or the chronology record.\n\n**CGS-2.8** A research pass, however deep, is a sweep and not a source; its findings are graded by the sources it cites, and its counts are verified against Grade A before use.\n\n---"
    },
    {
      "heading": "CGS-3 · Admissibility",
      "content": "**CGS-3.1** Evidence is admissible only for propositions the source directly supports. Source authority does not transfer to claims the source does not make.\n\n**CGS-3.2** Exploitation, actor identity, campaign linkage, motivation, and impact are separate propositions, each graded independently. Attribution does not inherit the grade of the exploitation it attaches to.\n\n**CGS-3.3** Where a criterion depends on event ordering, the registry records each date independently with its source: T_event, Evidence(T_event), for every event the criterion names. Inferred chronology cannot substitute for documented chronology unless the criterion explicitly permits inference. Number ranges, advisory serials, and adjacent evidence do not adjudicate a date.\n\n**CGS-3.4** Identifiers are matched in full form against admission fields only. A mention in a note is not an admission. A flag may be raised on Grade C or D; a miss may be declared only on Grade A dating.\n\n**CGS-3.5** An observation's founding instances must exist as signal rows before the observation opens.\n\n**CGS-3.6** Multi-identifier signals carry every identifier in the admission key.\n\n---"
    },
    {
      "heading": "CGS-4 · Criterion change",
      "content": "**CGS-4.1** Clarification: wording changes that alter no case's classification. Recorded; no version increment; no effective-date delay.\n\n**CGS-4.2** Refinement: a change to measurement that preserves the construct. Version increments; declared with an effective date no earlier than the next scheduled review.\n\n**CGS-4.3** Replacement: the prior criterion is found not to measure the construct. Version increments; the prior version is invalidated under CGS-8.\n\n**CGS-4.4** A revised criterion may not produce a rating action from the observation that caused its revision during the review in which the revision is declared. After the revised criterion becomes effective, the observation may be classified retrospectively for record completeness, but it does not independently satisfy a rating trigger unless the criterion expressly defines historical accrual.\n\n**CGS-4.5** Every substantive change records: version, declaration date, effective date, change type, reason, expected directional effect on qualifying observations, affected conditions, and retrospective test results per CGS-7.\n\n**CGS-4.6** No criterion is refined or replaced while an evaluation window it governs is open, except following invalidation under CGS-8. Invalidation takes effect immediately; a successor criterion follows CGS-4.5 and does not take effect before its declared effective date. Windows close on the scheduled review date.\n\n**CGS-4.7** A criterion for a state not yet reached, including retirement, is declared before the state is first reached, never at need.\n\n---"
    },
    {
      "heading": "CGS-5 · Exceptional review",
      "content": "**CGS-5.1** Review outside the schedule is permitted only on: a criterion found invalid; an authoritative source materially correcting underlying evidence; a duplicate observation discovered; a failure of evidence provenance; a state-machine ambiguity; or previously unavailable Grade A or B evidence that materially changes a case.\n\n**CGS-5.2** Exceptional review does not permit discretionary reclassification because evidence is interesting, surprising, or contrary to the board's thesis. The trigger is named in the record before the review begins.\n\n**CGS-5.3** Proximity to a trigger raises the evidence bar for the candidate that would meet it; it never lowers it.\n\n---"
    },
    {
      "heading": "CGS-6 · Conflict and independence",
      "content": "**CGS-6.1** When admissible sources conflict on a material fact, direct evidence of the proposition controls over indirect evidence regardless of general source grade. Where both sources contain direct evidence, the fact remains contested unless the discrepancy can be reconciled from the underlying record. Source grade alone does not resolve conflicting direct evidence: grade governs admissibility; evidentiary proximity governs conflict.\n\n**CGS-6.2** Multiple reports derived from one incident, vendor statement, telemetry set, or research document count as one evidentiary instance unless independently corroborated.\n\n**CGS-6.3** The counting unit for each condition is stated in its criterion. For the vendor risk-signal ledger the unit is the vendor assessment event. For exploitation conditions the unit is the sub-class instance. For containment and adversarial operations the unit is the distinct operation.\n\n**CGS-6.4** A sub-class is ratified on a definition with clean exclusions writable in one sentence, never on a shared noun, and never on count alone.\n\n**CGS-6.5** The taxonomy declares its gaps. An instance that fits no declared sub-class is recorded unattributed rather than forced.\n\n---"
    },
    {
      "heading": "CGS-7 · Retrospective testing",
      "content": "**CGS-7.1** Before activation, a criterion is tested against a set of historical observations covering known positives, known negatives, boundary cases, ambiguous cases, and adversarial cases.\n\n**CGS-7.1a** The test set contains all reasonably available qualifying historical cases when fewer than ten are known. Where ten or more cases are available, the test set contains at least ten and includes every category named in CGS-7.1 where such a case exists.\n\n**CGS-7.2** Results are published in three categories: QUALIFYING, NON-QUALIFYING, INDETERMINATE. Where the evidence cannot establish what the criterion needs, the case is indeterminate. Cases are not forced into yes or no.\n\n**CGS-7.3** The test set and results are recorded with the criterion version and are not revised after activation.\n\n---"
    },
    {
      "heading": "CGS-8 · Invalidation and review states",
      "content": "**CGS-8.1** A criterion is invalidated when its observable is found not to measure the construct it was designed to assess, or when a material defect makes its classifications unreliable.\n\n**CGS-8.2** An invalidated criterion produces no further rating actions.\n\n**CGS-8.3** A rating action whose sole qualifying basis was an invalidated criterion is VOIDED and the condition is placed UNDER REVIEW. The action is not reversed: reversal would decide the correct rating after the fact. The prior valid rating stands as the displayed state. The voided action, its date, and its criterion version remain visible in the record.\n\n**CGS-8.4** Every other adjudication made under the invalidated criterion is re-run under the valid successor at the same review.\n\n**CGS-8.5** Reclassification of a condition under review occurs only under a valid criterion at the next scheduled review or an exceptional review authorized under CGS-5.\n\n**CGS-8.6** The board displays a condition under review with: current displayed rating (the prior valid state), the voided action and its version, and the date and criterion version of the next adjudication.\n\n**CGS-8.7** A trigger met while a condition is at the top of the scale does not create a tier or a modifier. It is recorded, counted, and opens a Position-candidacy review under the issuance gate.\n\n**CGS-8.8** On de-escalation, prior adverse-event counts become historical context; re-escalation requires new events after the de-escalation.\n\n---"
    },
    {
      "heading": "CGS-9 · Intake and reconciliation",
      "content": "**CGS-9.1** Discovery runs against the authoritative data file, not against search or mirrors. Search and mirrors enrich; they do not discover.\n\n**CGS-9.2** Reconciliation runs source-to-registry across the full covered era, monthly, by full-form identifier.\n\n**CGS-9.3** The completeness-era boundary is a declared reconstruction. It is not moved to reclassify any absence.\n\n**CGS-9.4** Missed runs are declared gaps, never backdated. A gap obligates a source-file sweep of its window.\n\n**CGS-9.5** Every miss is classified, and every miss class carries a rule. The taxonomy is published.\n\n---"
    },
    {
      "heading": "CGS-10 · Publication",
      "content": "**CGS-10.1** Each issue cites the applied criteria by version and the provisions of this Standard under which each rating action was taken.\n\n**CGS-10.2** Board statistics carry rows for rating changes, scope refinements, watch changes, published triggers newly met with rating unchanged, rating actions voided, methodology changes, and corrections to prior issues. Rows are added when the machinery produces an event no row can hold; rows are never removed.\n\n**CGS-10.3** A correction to a prior issue is published at the volume of the original claim. An intake miss is disclosed with its class, its latency, and the rule that closes it.\n\n**CGS-10.4** Certainty is stated at the grade of the evidence. Reported is not confirmed; a label is not a chronology; a candidate is not a count.\n\n---"
    },
    {
      "heading": "CGS-11 · Governance of the Standard",
      "content": "**CGS-11.1** Changes to this Standard increment its version.\n\n**CGS-11.2** A CGS amendment records the provision changed, prior text, replacement text, reason, declaration date, effective date, and expected procedural effect.\n\n**CGS-11.3** No CGS amendment applies retroactively to alter the validity of an action taken under the Standard in force at the time.\n\n**CGS-11.4** An amendment arising from a live case cannot govern that case unless necessary to resolve a procedural state for which the Standard contains no rule. In that circumstance the amendment may govern procedure but cannot determine the substantive rating outcome.\n\n**CGS-11.5** Superseded versions remain permanently available.\n\n---"
    },
    {
      "heading": "Application on declaration",
      "content": "**SC-2026-006** is displayed under CGS-8.6 from this Standard's declaration: current rating STRENGTHENING (prior valid state); voided action CONFIRMED, 12 September 2026, criterion v1.0; next adjudication 24 September 2026 under criterion v2.0, whose CGS-1 validation and CGS-7 retrospective test are recorded in the registry before that date. Per CGS-4.4, the Fortinet observation that exposed v1.0 produces no rating action under v2.0 in the review that declared it; after v2.0 takes effect it is classified retrospectively for record completeness and does not independently satisfy a trigger.\n\n*End of Standard. CGS v1.0 is frozen. Provision identifiers are permanent.*"
    }
  ]
}
```