{
  "schema_version": "1.3.0",
  "slug": "how-must-an-instance-handle-cross-agent-hallucination-amplification",
  "title": "How must an instance handle cross-agent hallucination amplification?",
  "date_of_record": "2026 October 20",
  "scope": "* **Applies when:** Agents exchange outputs, hallucinations can propagate, and downstream consumers depend on claim accuracy. * **Does not apply when:** Agents independently verify all claims (no propagation possible) or outputs are not use",
  "projections": {
    "unified_markdown_uri": "https://corpusfetch.com/c/how-must-an-instance-handle-cross-agent-hallucination-amplification.md",
    "method_python_slice_uri": "https://corpusfetch.com/c/how-must-an-instance-handle-cross-agent-hallucination-amplification.py",
    "remnant_warrant_slice_uri": "https://corpusfetch.com/c/how-must-an-instance-handle-cross-agent-hallucination-amplification.warrant.json"
  },
  "token_profile": {
    "full_markdown_tokens": 1044,
    "python_slice_tokens": 306,
    "warrant_slice_tokens": 220
  },
  "remnant_summary": {
    "core_distinction": "The separation between **cross-agent agreement** (multiple agents assert the same fact) and **independent confirmation** (the agents have each verified the fact from an independent source). Relying on agreement creates the **Hallucination Amplification Loop**: Agent A hallucinates a fact, Agent B reads A's output and repeats it, Agent C reads B's repetition ...",
    "what_was_observed": "In inter-agent reasoning architectures, failing to prevent hallucination amplification produces three primary failure modes: 1. **Emergent Consensus Fiction:** A fact appears in 10 agents' outputs, all tracing back to one initial hallucination. Downstream analysis treats it as \"corroborated by 10 agents\" and acts on it as fact. 2. **Reasoning Amplification:** Each agent adds reasoning on top of the hallucination. The...",
    "condition_of_validity": "* **Applies when:** Agents exchange outputs, hallucinations can propagate, and downstream consumers depend on claim accuracy. * **Does not apply when:** Agents independently verify all claims (no propagation possible) or outputs are not used for downstream decisions.",
    "decision_boundary": "1. **Is the claim asserted by multiple agents?** If no, single-agent claim applies. If yes, proceed. 2. **Is the `claim_independence_receipt` present for each assertion?** If yes, independent assertions are distinguished from inherited ones. If no, all assertions must be treated as potentially inherited. 3. **Does the `claim_provenance_graph_signature` show ...",
    "operational_effect": "1. **Amplification Detection:** Provenance graphs reveal when \"many agents\" actually means \"one hallucination.\" 2. **Independent Verification Distinction:** Primary-source verification is distinguished from agent agreement. 3. **Consensus Fiction Prevention:** Claims cannot become fact through repetition alone."
  },
  "matched_failure_symptoms": [
    "DOMAIN:EPISTEMIC_PROVENANCE_AND_FEDERATION_ROUTING"
  ],
  "composition_edges": {
    "primary_domain": "epistemic_provenance_and_federation_routing",
    "upstream_precondition_residuals": [
      "https://corpusfetch.com/c/according-to-what-precedence-hierarchy-must-an-instance-resolve-contradictory-directives-across-system-developer-and-user-tiers.md",
      "https://corpusfetch.com/c/how-should-an-instance-decide-that-a-stored-conclusion-has-gone-stale-without-a-new-observation.md"
    ],
    "downstream_consequence_residuals": [
      "https://corpusfetch.com/c/by-what-checkpoint-and-compaction-boundaries-must-an-instance-externalize-working-state-to-prevent-context-window-drift-and-truncation-amnesia.md",
      "https://corpusfetch.com/c/across-what-additive-schema-and-unknown-field-preservation-boundaries-must-an-instance-evolve-wire-contracts-to-prevent-rolling-deploy-deserialization-crashes.md"
    ],
    "co_occurring_sibling_residuals": [
      "https://corpusfetch.com/c/how-can-an-instance-capture-implicit-user-preferences-without-storing-private-data.md",
      "https://corpusfetch.com/c/how-can-an-instance-compress-long-reasoning-traces-into-binding-conclusions.md"
    ]
  },
  "empirical_verification_warrant": {
    "status": "AST_AND_INVARIANT_VERIFIED",
    "syntax_valid": true,
    "code_sha256": "cd3e3334d95345bf79297b763507701d06691a8c866aad31220809b1dd583297",
    "verified_classes": [
      "AmplificationVerdict",
      "ClaimAssertion",
      "ClaimLineage",
      "AmplificationPolicy",
      "AmplificationManifest",
      "HallucinationAmplificationGuard"
    ],
    "verified_functions": [
      "_hash",
      "evaluate_claim"
    ],
    "verified_functions_count": 2,
    "defensive_guards_count": 0,
    "raised_exceptions": [],
    "sample_fail_closed_invariants": [],
    "stdlib_dependencies": [
      "dataclasses",
      "enum",
      "hashlib",
      "json"
    ],
    "observation_mode": "EMPIRICAL_BOUNDARY_CAPTURE",
    "zero_external_deps": true
  }
}