{
  "scope": "Exposition-to-evidence map, not a Lean theorem ledger.",
  "claims": [
    {
      "id": "KME01",
      "location": "2: question sufficiency",
      "claim": "Factorization through the image iff constancy on representation fibers.",
      "evidence_kind": "written proof",
      "current_support": "Finite set tests corroborate instances.",
      "boundary": "Sets/functions on the declared domain, no computability claim."
    },
    {
      "id": "KME02",
      "location": "2: continuation",
      "claim": "Represented dynamics factor iff represented successors are constant on fibers; iteration follows by induction.",
      "evidence_kind": "written proof",
      "current_support": "216 finite representation/transition scenarios.",
      "boundary": "Unrestricted state claim distinguished from restricted initial states."
    },
    {
      "id": "KME03",
      "location": "2: coherent views",
      "claim": "Every finite prefix of the inclusion tower has a section but no global compatible section exists.",
      "evidence_kind": "written proof",
      "current_support": "Finite prefixes illustrate but do not prove the infinite claim.",
      "boundary": "Specified natural-number fibers and inclusion maps."
    },
    {
      "id": "KME04",
      "location": "3: returning coordinate",
      "claim": "PT²P=(PTP)²+PTQTP for complementary projectors.",
      "evidence_kind": "written proof",
      "current_support": "Rational example and seeded finite matrix checks.",
      "boundary": "Linear operators with P²=P and Q=I−P."
    },
    {
      "id": "KME05",
      "location": "3: memory formula, equation 2",
      "claim": "Block elimination gives exact history terms and an initial-complement term.",
      "evidence_kind": "written induction",
      "current_support": "Direct block trajectories compared to memory trajectories, including perturbations and nonzero initial state.",
      "boundary": "Finite sums at each time; no convergence or numerical-efficiency claim."
    },
    {
      "id": "KME06",
      "location": "3: defects, equation 3",
      "claim": "Composite defect equals the two correctly transported defects.",
      "evidence_kind": "written proof",
      "current_support": "Typed finite matrix checks.",
      "boundary": "Linear additive setting with matching maps."
    },
    {
      "id": "KME07",
      "location": "4: correction",
      "claim": "Correction exists iff the defect class vanishes in the declared quotient image.",
      "evidence_kind": "written proof",
      "current_support": "Diagonal obstruction functional and enlarged-space solution.",
      "boundary": "Protected correction uses D(ker p), not unrestricted image."
    },
    {
      "id": "KME08",
      "location": "4: stability",
      "claim": "A vanishing residual need not imply distance to a true solution tends to zero.",
      "evidence_kind": "written proof",
      "current_support": "Unit-coordinate finite checks are illustrative.",
      "boundary": "Bounded injective diagonal operator on ℓ²; no general stability theorem claimed."
    },
    {
      "id": "KME09",
      "location": "5: arithmetic",
      "claim": "Euler parity transform and coefficientwise roots-of-unity selector.",
      "evidence_kind": "written symbolic proof",
      "current_support": "Exact integer/rational coefficient checks and an exact finite root specialization.",
      "boundary": "No automatic integral or p-adic divisibility consequence."
    },
    {
      "id": "KME10",
      "location": "6: verification",
      "claim": "Kernel acceptance, axiom policy, formal meaning, ordinary hypotheses, and reproduction are distinct.",
      "evidence_kind": "official documentation and methodological explanation",
      "current_support": "Current official-source consultation; historical project example attributed.",
      "boundary": "No new Lean execution or independent kernel."
    },
    {
      "id": "KME11",
      "location": "7: proof fields",
      "claim": "AND-premise dependencies, alternative justifications, and candidate edges have distinct roles.",
      "evidence_kind": "defined finite illustration and cited project theory",
      "current_support": "Finite certificate-closure checks and negative controls.",
      "boundary": "Not a model of every human thought or recovered discovery history."
    },
    {
      "id": "KME12",
      "location": "8: collaboration",
      "claim": "Inspectable records may improve handoff and reuse.",
      "evidence_kind": "methodological proposal",
      "current_support": "Operator case gives a concrete proposed handoff.",
      "boundary": "No measured productivity advantage asserted."
    },
    {
      "id": "KME13",
      "location": "9: evaluation",
      "claim": "Compare structured records with ordinary notes under declared tasks, budgets, and review.",
      "evidence_kind": "proposed empirical protocol",
      "current_support": "No comparative experiment conducted.",
      "boundary": "Effectiveness remains open to testing and possible adverse results."
    },
    {
      "id": "KME14",
      "location": "source package",
      "claim": "Complete-inventory verifier rejects targeted invalid archives.",
      "evidence_kind": "executed implementation tests",
      "current_support": "17 actual CLI fixtures with intended diagnostics.",
      "boundary": "Integrity relative to manifest, not authentication or mathematical soundness."
    }
  ]
}