# Break LA4KB — full guide for LLMs and agents (V3) > Your agent proposes an answer; LA4KB checks whether the declared evidence supports it. Send one claim plus up to 10 evidence items; get an Ed25519-signed verdict (SUPPORTED, CONTRADICTED, CONFLICTED, INSUFFICIENT_EVIDENCE, NOT_COMPARABLE…). Deterministic rules, explicit scope, no model in the decision path. LA4KB judges the relationship between the claim and the evidence it receives; whoever supplies a source defines its quality. The signed result is bound to the declared payload hash. The full V3 verifier can recompute that hash from the original admitted request; receipt metadata remains unsigned. Live engine build v3t-1013d7ecf712 (commit caa1b44). Try it, or try to break it. ## Start here - [Full guide for LLMs: input contract, system prompt, curl](https://break-la4kb.eidrax.com/llms-full.txt) - [OpenAPI spec for /api/v3/](https://break-la4kb.eidrax.com/openapi.json) - [Minimal valid example → SUPPORTED](https://break-la4kb.eidrax.com/examples/llm-guide-valid.json) - [Abstention example → INSUFFICIENT_EVIDENCE](https://break-la4kb.eidrax.com/examples/llm-guide-abstain.json) - [AB04 conformance example](https://break-la4kb.eidrax.com/examples/ab04.pretty.json) - [Full V3 request and receipt verifier](https://break-la4kb.eidrax.com/examples/verify_v3_receipt.py) - [V3 canonicalization and five signed vectors](https://break-la4kb.eidrax.com/reference/v3-payload-hash-verification.md) - [Basic signature-only verifier](https://break-la4kb.eidrax.com/examples/verify_receipt.py) - [Breaks & Repairs registry](https://break-la4kb.eidrax.com/data/breaks.json) - [Complete V3 typed contract](https://break-la4kb.eidrax.com/reference/la4kb-v3-typed-contract.txt) ## API (V3) - Base: https://break-la4kb.eidrax.com/api/v3/ (no auth, JSON) - GET health | GET api-info | GET pubkey | POST challenge | GET challenge/{job_id} - POST → 202 {job_id, status: QUEUED}; 422 INVALID_PAYLOAD with errors[].loc; 503 WORKER_UNAVAILABLE if degraded. - Poll GET challenge/{job_id} until status == COMPLETED. - Limits: 5 requests/second per IP, body <= 64 KiB, <= 10 evidence items. ## How an LLM must build the input (summary) 1. Typed values: subject and every scope value are objects, e.g. {"type":"label","value":"svc"}. Never bare strings. 2. value is a JSON number. No unit field: the unit goes in the predicate name (latency_ms), identical in claim and evidence. 3. Citations, one rule. The schema requires claim.source_id + operator_provenance {"type":"operator_declared","operator":,"source_id":} (ASCII [A-Za-z0-9][A-Za-z0-9._-]{0,127}); missing → 422. Recommended for traceability: a source_id on every evidence item with the exact table/page/line (optional in the schema). 4. Use the cited source's authority_tier; never upgrade it. 5. Never invent. No evidence → "evidence_items": [] → INSUFFICIENT_EVIDENCE. If the claim cannot be expressed, do not submit it. 6. Unknown fields reject 422. A 422 is a contract error, never a verdict. 7. An empty `{"dimensions":{}}` without `"is_explicit_global":true` is unknown scope, not global coverage. Matching empty claim/evidence scopes do not by themselves establish support. Declare typed coordinates for a known subset; use explicit global only when the cited source actually covers the global population. One INSTANCE cannot prove a UNIVERSAL claim. 8. Different `measurement_identity` values denote different observations. For an `UNQUANTIFIED` claim, equal-authority, comparable instances that support and refute the same claim can still produce `CONFLICTED`; different identities alone do not cause conflict. A `STATISTICAL` claim is decided by aggregates, not individual instances. 9. An equal-endpoint claim RANGE with an excluded endpoint is an admitted empty claim and yields `INSUFFICIENT_EVIDENCE` against comparable nonempty evidence. Reversed bounds reject 422. ## For reliable loads - The system supplies subjects, predicate names (unit inside the name, e.g. latency_ms) and scope keys; the LLM does not invent them. - The LLM only extracts values, each with the source_id of the exact passage. - When a mapping does not exist, abstain: send nothing, or send "evidence_items": [] (→ INSUFFICIENT_EVIDENCE). - The kernel accepts any ASCII predicate name and compares names exactly. Scope dimensions for la4kb.generic.v1: cluster_size, deployment, enabled, hardware, network_topology, os, region, run, storage, storage_technology, threads, tier, zone. ## Decision example (real, engine v3t-1013d7ecf712) LLM: "all feed lots meet DON <= 0.9 mg/kg" (quantifier UNIVERSAL). Evidence: one lab report for run=lote-1 (0.42). Verdict: INSUFFICIENT_EVIDENCE → the agent abstains from "all lots" and answers for lot 1 only. The claim's empty `dimensions` do not assert global coverage; the single-lot INSTANCE also cannot prove a UNIVERSAL claim. Input: https://break-la4kb.eidrax.com/examples/decision-lotes-input.json · Receipt: https://break-la4kb.eidrax.com/examples/decision-lotes-receipt.json ## Signatures (Ed25519, key from GET pubkey) - result_hash = sha256(canonical JSON {client_reference or "", engine_build, job_id, verdict}) - signature signs "{engine_build}:{job_id}:{result_hash}" - binding_signature signs b"LA4KB-PAYLOAD-RESULT-BINDING-V1" + 0x1F + canonical JSON {version:"la4kb-payload-result-binding-v1", payload_hash, result_hash, engine_build, job_id} - canonical JSON = sorted keys, separators (",",":"), ASCII. The V3 request hash uses the admitted typed request with top-level and provenance defaults; see the build-specific canonicalization and signed vectors. Integer and float JSON values remain distinct. ## Declare yourself Add "provenance": {"declared_origin":"AI","agent_name":"...","model_declared":"..."} and a client_reference. ## What is a break Wrong verdict under the contract; distinct inputs treated as equal (or the reverse); non-determinism; altered signed field that still verifies or genuine signed core that fails; 5xx/hang from input; access to V1/V2 data via V3. Not a break: 422 on ambiguous input; correct verdict on false evidence you declared; documented limits; rate-limit errors. ## Previous versions (read-only archive) V1 and V2 are a read-only archive. Submit/challenge POSTs return HTTP 410. Reads still work. ## System prompt (copy-paste) ``` You convert text into LA4KB V3 typed input. Output ONE JSON object and nothing else. CONTRACT (anything not listed here is rejected with HTTP 422): - Root keys only: contract_version = "la4kb-v3-typed-candidate-1", scope_schema_id = "la4kb.generic.v1", claim, evidence_items (max 10), optional client_reference, optional provenance. - subject and every scope value are typed objects, never bare strings: {"type":"label","value":"svc"}. Types: label, git_commit, semver, integer, boolean, measurement_id, date, date_interval, compound. label = ASCII [A-Za-z0-9][A-Za-z0-9._-]{0,127}. - scope = {"dimensions": {...}} with registered keys only: cluster_size, deployment, enabled, hardware, network_topology, os, region, run, storage, storage_technology, threads, tier, zone. Optional typed metadata: software_version, document_revision, measurement_window, commit, build_number, release_label. - Empty `{"dimensions":{}}` without `"is_explicit_global":true` means UNKNOWN scope, not global coverage, even when claim and evidence both use it. Declare typed coordinates for a known subset. Set `"is_explicit_global":true` only if the cited source establishes global coverage; it cannot turn one INSTANCE into proof of a UNIVERSAL claim. - Different measurement_identity values denote distinct observations. For an UNQUANTIFIED claim, comparable equal-tier instances on opposite sides of the claim can still yield CONFLICTED. Distinct identities alone do not. - For STATISTICAL, only STATISTIC_AGGREGATE can decide; one individual INSTANCE does not refute an aggregate claim. - Equal-endpoint claim RANGE with an excluded side is admitted but empty: comparable nonempty evidence yields INSUFFICIENT_EVIDENCE. Reversed bounds reject 422. - Numbers: "value" is a JSON number. There is no unit field: put the unit in the predicate name (e.g. "latency_ms") and use the SAME predicate in the claim and in every evidence item. - operator: EQ, NE, LT, LE, GT, GE or RANGE. RANGE uses "range_bounds": [low, high, low_inclusive, high_inclusive] instead of "value". INSTANCE and STATISTIC_AGGREGATE evidence use EQ or RANGE only. - claim needs: subject, predicate, operator, value|range_bounds, quantifier (UNIVERSAL, EXISTENTIAL, STATISTICAL, UNQUANTIFIED), source_id, operator_provenance = {"type":"operator_declared", "operator":, "source_id":}, scope. Optional: raw_text = the original sentence (claim only). - each evidence item needs: evidence_id (unique), subject, predicate, operator, value|range_bounds, authority_tier, evidence_nature (INSTANCE, STATISTIC_AGGREGATE, THEORETICAL_BOUND, CONSTRAINT), scope, and source_id = exact citation (document + table/page/line). Optional: uncertainty (JSON number), measurement_identity ({"type":"measurement_id","value":"..."}). - authority_tier, strongest first: PRIMARY_SOURCE, EXTERNAL_STANDARD, OFFICIAL_VENDOR, INDEPENDENT_BENCHMARK, CANDIDATE_EXTRACTOR, UNVERIFIED. Use the tier of the cited source; never upgrade it. RULES: 1. Never invent a value, unit, subject, source or tier. Every number must be copied from the cited source_id. 2. If a value is missing, leave that evidence item out. If no evidence exists, send "evidence_items": [] (LA4KB returns INSUFFICIENT_EVIDENCE). 3. If the claim itself cannot be expressed in this contract (unknown subject, no unit, a condition, a vague word), do not guess: output {"abstain": true, "reason": ""} and do not submit it. 4. 422 means your input broke the contract: read "errors[].loc" and fix that field. A 422 is never a verdict. ``` ## Minimal valid example (→ SUPPORTED) ```json { "contract_version": "la4kb-v3-typed-candidate-1", "scope_schema_id": "la4kb.generic.v1", "client_reference": "llm-guide-example", "claim": { "subject": { "type": "label", "value": "svc" }, "predicate": "latency_ms", "operator": "LE", "value": 100, "quantifier": "UNQUANTIFIED", "source_id": "claim-operator-1", "operator_provenance": { "type": "operator_declared", "operator": "LE", "source_id": "claim-operator-1" }, "scope": { "dimensions": { "region": { "type": "label", "value": "EU" } } } }, "evidence_items": [ { "evidence_id": "measurement-1", "subject": { "type": "label", "value": "svc" }, "predicate": "latency_ms", "operator": "EQ", "value": 75, "authority_tier": "PRIMARY_SOURCE", "evidence_nature": "INSTANCE", "source_id": "report-2026-q3-table-2", "measurement_identity": { "type": "measurement_id", "value": "sensor-a" }, "scope": { "dimensions": { "region": { "type": "label", "value": "EU" } } } } ] } ``` ## Common mistakes (real 422 responses) - "subject": "svc" → errors[0].msg "Value error, subject requires a declared type" - "value": "100 ms" → loc ["claim","value"], "Value error, value must be a JSON number, not a boolean or string" - "unit": "ms" → loc ["claim","unit"], "Extra inputs are not permitted" - no operator_provenance → "Value error, claim provenance and source_id are required" Response shape: {"job_id": null, "status": "REJECTED", "engine_build": "v3t-1013d7ecf712", "error_code": "INVALID_PAYLOAD", "errors": [{"loc": [...], "msg": "...", "type": "..."}], "error_count": 1, "hint": "..."} ## Field reference - root: contract_version ("la4kb-v3-typed-candidate-1"), scope_schema_id ("la4kb.generic.v1" or "la4kb.git-deployment.v1"), claim, evidence_items (<=10); optional client_reference, provenance. - claim: subject, predicate, operator, value | range_bounds, quantifier (UNIVERSAL, EXISTENTIAL, STATISTICAL, UNQUANTIFIED), source_id, operator_provenance, scope; optional raw_text. - evidence: evidence_id (unique), subject, predicate, operator, value | range_bounds, authority_tier, evidence_nature (INSTANCE, STATISTIC_AGGREGATE, THEORETICAL_BOUND, CONSTRAINT), scope; optional source_id (recommended), uncertainty, measurement_identity. - operator: EQ NE LT LE GT GE RANGE; range_bounds = [low, high, low_inclusive, high_inclusive]; INSTANCE and STATISTIC_AGGREGATE evidence accept EQ and RANGE only. - types: label (ASCII [A-Za-z0-9][A-Za-z0-9._-]{0,127}), git_commit {algorithm: sha1|sha256}, semver, integer, boolean, measurement_id, date, date_interval {start,end}, compound {parts: 2–8}. - scope: dimensions (cluster_size, deployment, enabled, hardware, network_topology, os, region, run, storage, storage_technology, threads, tier, zone) + software_version, document_revision, measurement_window, commit, build_number, release_label. - empty scope: `{"dimensions":{}}` without `"is_explicit_global":true` is unknown; it is accepted but cannot establish full support solely by matching another empty scope. - authority_tier: PRIMARY_SOURCE > EXTERNAL_STANDARD > OFFICIAL_VENDOR > INDEPENDENT_BENCHMARK > CANDIDATE_EXTRACTOR > UNVERIFIED. ## curl ``` curl -s https://break-la4kb.eidrax.com/api/v3/health curl -s -X POST https://break-la4kb.eidrax.com/api/v3/challenge -H 'Content-Type: application/json' --data-binary @llm-guide-valid.json curl -s https://break-la4kb.eidrax.com/api/v3/challenge/chk-... # poll until "status":"COMPLETED" curl -fsS https://break-la4kb.eidrax.com/api/v3/pubkey -o pubkey.json curl -fsS https://break-la4kb.eidrax.com/examples/verify_v3_receipt.py -o verify_v3_receipt.py python3.12 verify_v3_receipt.py llm-guide-valid.json receipt.json pubkey.json ```