Skip to content

Evidence pack

An evidence_pack is the portable container for all evidence facts related to one scoped agent outcome. It can represent an answer, artifact, task, run, session, review, incident, support case, or external handoff.

Required fields

FieldPurpose
evidence_pack_idStable pack id.
schema_versionAgent Evidence schema version.
scopeSession, task, run, artifact, answer, review, incident, or external case scope.
statusPack lifecycle status.
created_at / updated_atProducer timestamps.
producerRuntime, service, worker, host, or exporter that assembled the pack.
claims / claim_map_refInline compact claims or a referenced claim map.
sources / source_map_refInline compact sources or a referenced source map.
support_edgesRelationships among claims, sources, checks, artifacts, and provenance.
provenance / provenance_refProduction graph or compact provenance refs.
verification_resultsCheck results.
reviewsHuman, automated, or policy verdicts.
replay_casesReplay and reconstruction boundaries.
redactions / redaction_summaryRedaction and access facts.
completenessCategory-level completeness.

A pack SHOULD prefer refs for large payloads. It MAY embed compact facts needed for offline audit.

Completeness entry

Each completeness category SHOULD state:

FieldPurpose
statuscomplete, partial, missing, unknown, not_applicable, or not_collected.
missing_factsStructured missing fact records.
notesOptional human-readable explanation.
last_checked_atTimestamp for freshness-sensitive categories.

Pack invariants

  • A pack MUST keep ids stable across redaction and export.
  • A pack MUST NOT mark missing evidence as passed.
  • A pack SHOULD preserve the relationship graph even when snippets or payloads are removed.
  • A pack SHOULD disclose whether it is a full pack, compact projection, redacted export, or pointer-only handoff.
  • A pack SHOULD include schema refs and hashes when exported outside the source system.

Draft standard for portable agent evidence, provenance, review, and replay.