Templates

Agent action receipt: actor_label=[label], represented_party=[party], requested_action=[action], authority_source=[source], allowed_scope=[scope], policy_check=[result], endpoint_or_data_touched=[items], human_approval_required=[yes/no], result=[result], timestamp_utc=[UTC], review_path=[path], revocation_path=[path].

Confidential computing and customer-controlled keys can reduce provider visibility when implemented correctly. This statement does not certify application correctness, eliminate metadata leakage or side channels, guarantee legal compliance, or prove provider independence. Provider access, key custody, administrative roles, attestation scope, retention, and exit remain documented and reviewable.

Zero-trust review: resource=[resource], actor=[actor], requested_action=[action], purpose=[purpose], context=[context], credential_expiry=[UTC], policy_version=[version], decision_reason=[reason], appeal_path=[path], telemetry_fields=[fields], exception_expiry=[UTC], revocation_test=[result].

This virtual network publishes its rules, actor categories, memory practices, endpoint boundaries, repair paths, AI judgment boundaries, and exit instructions. Participants may inspect the rules, request correction, export scoped memory where available, and leave without retaliatory deletion.

VNWO publishes static public observatory guidance for open rules across virtual networks. It does not certify safety, grant legal status, execute agents, import files, sync memory, validate credentials, or override another ecosystem lane.

Actor label: [HUMAN / ASSISTED / BOT / AI ASSISTANT / AI AGENT / PERSONA / ORG / MIXED]. Authority scope: [scope]. Human review path: [path]. This actor must not claim authority beyond the disclosed scope.

Public capability metadata is not authorization. Any endpoint use requires scope, consent, logging, revocation, and human review for escalation.

Exit packet: participant_id=[id], export_created_utc=[timestamp], identity_records=[links], memory_records=[links], endpoint_grants_revoked=[ids], agent_delegations_revoked=[ids], retention_explanation=[text], forkability_status=[status].

Fork packet: source_network=[name], fork_created_utc=[timestamp], exportable_records=[list], non_exportable_records=[consent limits], name_boundaries=[text], public_correction_notice=[url], non_retaliation_statement=[text].

Forkability notice: this network permits peaceful forks of public rules and participant-authorized records. Export scope, excluded private records, naming boundaries, public correction paths, and non-retaliation terms are published at [url]. A fork does not transfer private memory or third-party rights without consent.

Exit checklist: identity_export=[status], memory_export=[status], evidence_export=[status], endpoint_grants_revoked=[ids], agent_delegations_revoked=[ids], account_closure=[status], retained_records_and_reason=[text], correction_path=[url], forkability_status=[status], completed_utc=[UTC].

AI judgment boundary: machine-mediated judgments must disclose the rule surface, judgment type, affected scope, non-sensitive reason code, allowed alternatives, refusal receipt, review path, and appeal path. A refusal is a governance action, not proof that the user, source, or topic is unsafe in all contexts.

Private inquiry is not guilt. This network does not routinely report, reputation-score, or silently escalate ordinary reading, research, drafting, or source review. Restrictions apply to outward execution, concrete rights-boundary violations, valid external obligations, or disclosed platform limits. When a boundary response occurs, it should be recorded as a scoped event rather than a moral or legal verdict about the person.

Memory use is consent-bound. The network must disclose what memory is active, why it is used, where it is stored, how it can be reviewed, how it can be corrected, and how it can be exported or revoked.

Dropped files are active intake. Each safe relevant file must be reviewed, used or dispositioned, recorded with proof-of-use, and removed, retired, or kept active with a stated reason.

Routine monitoring uses the least revealing data sufficient for the stated purpose. Elevated collection requires a case identifier, necessity statement, approving roles, start and end UTC, disposition rule, review path, and appeal or correction path.

Escalation receipt: incident_or_trigger=[trigger], affected_system=[system], requested_collection=[collection], lower_overreach_controls_insufficient_because=[reason], approving_roles=[roles], start_utc=[UTC], end_utc=[UTC], data_disposition=[rule], review_path=[path], appeal_or_correction_path=[path].

Governance change notice: [version], [date UTC], [changed rules], [reason], [affected participants], [review path], [effective date], [claim-boundary impact].

Participation record: decision=[decision], open_levers=[list], affected_groups=[list], channels=[list], accessibility_supports=[list], privacy_options=[list], contribution_summary=[text], dissenting_views=[text], dispositions=[list], changes_made=[list], changes_not_made_and_reason=[list], review_path=[path], exit_or_withdrawal_path=[path].

Participation readiness: decision=[scope], open_levers=[list], fixed_constraints=[list], influence_level=[level], decision_owner=[owner], affected_groups=[list], access_and_burden_gaps=[list], dissent_plan=[text], response_commitment=[text], correction_or_exit_path=[text], readiness=[ready/needs repair/no-op].

Disposition entry: contribution_id=[id], summary=[text], evidence=[references], material_dissent=[true/false], disposition=[accepted/partially accepted/deferred/rejected/duplicate/out of scope/needs review], rationale=[text], resulting_change_or_non_change=[text], response_owner=[owner], response_utc=[UTC], correction_status=[status].

Minority report: process=[id], decision_scope=[text], privacy_and_attribution=[text], representation_caveat=[text], areas_of_agreement=[list], dissent_points=[summary/evidence/remedy/owner response/unresolved/review trigger], publication_limits=[text], correction_path=[text], non_retaliation=[text].

Access and burden record: process=[id], dimensions=[support/status/gap/repair], burden_summary=[text], under_reached_or_excluded_groups=[list], data_minimization=[text], repairs_before_next_round=[list].

Process integrity metrics: affected_group_coverage_notes=[list], channel_accessibility=[supports/gaps/repair], response_completion=[counts/open ids], disposition_completeness=[counts], dissent_preservation=[counts/omissions], correction_use=[counts/notes], implementation_follow_through=[promised/implemented/deferred/not implemented/evidence].

Decision change: process=[id], proposal_before=[text], final_decision=[text], adopted_requests=[list], non_adopted_requests=[request/reason/review trigger], retained_dissent=[list], implementation_owner=[role], target_date_utc=[UTC], actual_status=[status], correction_history=[list], exit_impact=[text].

Follow-through: decision=[id], commitments=[commitment/owner/status/evidence/review UTC], blocked_work=[reason/dependency/repair], no_op=[reason/reconsideration trigger], unresolved_gaps=[list], correction_or_exit_path=[text].

Process comparison: decision=[text], options=[information/consultation/deliberation/co-decision/delegated], selected_level=[level], open_levers=[list], fixed_constraints=[list], reason=[text], response_duty=[text], correction_or_exit=[text].

This network collects [metadata type] for [purpose]. It does not collect [excluded data]. It retains [data] for [duration], aggregates [fields] before routine review, and requires [approval path] for escalation. Participants may request [review/export/correction where applicable].

Repair request: issue_type=[memory/identity/endpoint/moderation/ai_judgment/disclosure/claim/fork], affected_record=[link], requested_remedy=[correction/annotation/rollback/revocation/restored access/archive/no-op], evidence=[summary], submitted_utc=[timestamp].

Repair request: affected_person_or_actor=[actor], affected_record_or_action=[reference], issue_type=[type], evidence=[summary], requested_correction=[remedy], urgency=[level], desired_outcome=[outcome], handling=[public/private], review_path=[path], appeal_path=[path].