{"ok":true,"endpoint":"frontier-index/catalog","service":"a11oy.frontier.index","title":"Frontier Index — honest ecosystem catalog + self-audit","label":"MODELED","what":"an honest, self-auditing enumeration of every registered frontier surface with the data label its OWN backend actually emits, its cited paper(s), and whether it is a11oy-native or a cross-origin fallback. Derived live from the app's surface registry + registered routes + each surface's own response — never a hand-maintained list that can drift.","introspection":{"surface_registry":"szl3d_holographic.SURFACES (imported in-process)","route_source":"app.routes under /api/a11oy/v1 (GET)","label_source":"each surface's own in-process response (VERBATIM, never upgraded)","api_routes_seen":914},"doctrine":{"label_top":"MODELED","locked_proven":8,"locked_set":["F1","F4","F7","F11","F12","F18","F19","F22"],"kernel_commit":"c7c0ba17","adds_to_locked_8":0,"lambda":"Conjecture 1","khipu_bft":"Conjecture 2","trust_ceiling":0.97,"trust_100_percent":false,"runtime_cdn":0,"note":"additive read-only surface; touches no locked formula and no kernel; signs/mints nothing on a GET; introduces no theorem, no green/1.0."},"labels_legend":{"a11oy-native":"a local /api route for this surface answered in-process with an honest label","cross-origin-fallback":"route registered but no a11oy-native honest label emitted (proxy/other origin)","frontend-only":"no local /api route; the surface renders client-side (MODELED/cross-origin)"},"honest_labels_vocabulary":["LIVE","MEASURED","MODELED","SAMPLE","SIMULATED","CACHED","PROVEN","CONJECTURE","DECLARED","ROADMAP","DEGRADED","REPLAY","STRUCTURAL-ONLY","HONEST-STUB","UNSIGNED-LOCAL","UNAVAILABLE"],"summary":{"surfaces":138,"backend_counts":{"frontend-only":48,"a11oy-native":90},"label_counts":{"UNAVAILABLE":74,"MODELED":56,"DECLARED":1,"ROADMAP":2,"MEASURED":2,"STRUCTURAL-ONLY":1,"LIVE":1,"PROVEN":1},"surfaces_with_citations":25},"surfaces":[{"id":"integritycontrol","title":"Integrity Control Plane","category":"governance","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"atlas","title":"Atlas","category":"map","backend":"a11oy-native","label":"MODELED","citations":[],"endpoint":"/api/a11oy/v1/atlas/manifest","routes_registered":4},{"id":"loopforge","title":"Loop Forge","category":"brain","backend":"a11oy-native","label":"MODELED","citations":[],"endpoint":"/api/a11oy/v1/loopforge/manifest","routes_registered":6},{"id":"frontier","title":"Frontier","category":"map","backend":"a11oy-native","label":"DECLARED","citations":[],"endpoint":"/api/a11oy/v1/frontier/projects/info","routes_registered":30},{"id":"neuromorphic","title":"Neuromorphic","category":"more","backend":"a11oy-native","label":"MODELED","citations":["arXiv:1901.09948","10.1109/JPROC.2021.3067593"],"endpoint":"/api/a11oy/v1/neuromorphic/spikes","routes_registered":1},{"id":"interpretability","title":"Interpretability","category":"governance","backend":"a11oy-native","label":"MODELED","citations":[],"endpoint":"/api/a11oy/v1/interpretability/features","routes_registered":1},{"id":"circuits","title":"Attribution Graph · Circuit Tracing","category":"governance","backend":"a11oy-native","label":"MODELED","citations":[],"endpoint":"/api/a11oy/v1/circuits/attribution","routes_registered":1},{"id":"worldmodel","title":"World Model","category":"attention","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"qec","title":"Topological QEC","category":"proof","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"episodic","title":"Episodic Memory","category":"brain","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"testtime","title":"Test-Time Compute","category":"reasoning","backend":"a11oy-native","label":"MODELED","citations":[],"endpoint":"/api/a11oy/v1/testtime/scaling","routes_registered":1},{"id":"ssm","title":"State-Space (Mamba-3)","category":"attention","backend":"a11oy-native","label":"UNAVAILABLE","citations":[],"endpoint":"/api/a11oy/v1/govern/manifest/ssm","routes_registered":1},{"id":"genie","title":"Genie World-Model","category":"attention","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"specdecode","title":"Speculative Decoding","category":"attention","backend":"a11oy-native","label":"UNAVAILABLE","citations":[],"endpoint":"/api/a11oy/v1/govern/manifest/specdecode","routes_registered":1},{"id":"flowmatch","title":"Flow Matching","category":"attention","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"dllm","title":"Diffusion LLM","category":"attention","backend":"a11oy-native","label":"MODELED","citations":[],"endpoint":"/api/a11oy/v1/dllm/denoise","routes_registered":1},{"id":"moe","title":"MoE Router","category":"attention","backend":"a11oy-native","label":"UNAVAILABLE","citations":[],"endpoint":"/api/a11oy/v1/govern/manifest/moe","routes_registered":1},{"id":"formalmath","title":"Formal-Math Retrieval","category":"governance","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"ccattest","title":"Confidential-Compute Attest","category":"proof","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"attestinfer","title":"Attested Inference","category":"proof","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"pcai","title":"Proof-Carrying Attested Inference","category":"proof","backend":"a11oy-native","label":"MODELED","citations":[],"endpoint":"/api/a11oy/v1/pcai/run","routes_registered":1},{"id":"ringattn","title":"Ring Attention","category":"attention","backend":"a11oy-native","label":"UNAVAILABLE","citations":[],"endpoint":"/api/a11oy/v1/govern/manifest/ringattn","routes_registered":1},{"id":"grpo","title":"GRPO Reward Dynamics","category":"rl","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"kvcache","title":"KV-Cache H2O","category":"quant","backend":"a11oy-native","label":"ROADMAP","citations":[],"endpoint":"/api/a11oy/v1/energy/kvcache","routes_registered":2},{"id":"mla","title":"Latent Attention (MLA)","category":"attention","backend":"a11oy-native","label":"MODELED","citations":[],"endpoint":"/api/a11oy/v1/mla/latent-compress","routes_registered":1},{"id":"blt","title":"Byte-Latent Patching","category":"attention","backend":"a11oy-native","label":"UNAVAILABLE","citations":[],"endpoint":"/api/a11oy/v1/govern/manifest/blt","routes_registered":1},{"id":"nsa","title":"Native Sparse Attention","category":"attention","backend":"a11oy-native","label":"UNAVAILABLE","citations":[],"endpoint":"/api/a11oy/v1/govern/manifest/nsa","routes_registered":1},{"id":"kan","title":"Kolmogorov-Arnold Network","category":"attention","backend":"a11oy-native","label":"MODELED","citations":["arXiv:2404.19756"],"endpoint":"/api/a11oy/v1/kan/fit","routes_registered":1},{"id":"steering","title":"Steering Vectors (ActAdd / RepE)","category":"governance","backend":"a11oy-native","label":"UNAVAILABLE","citations":[],"endpoint":"/api/a11oy/v1/govern/manifest/steering","routes_registered":1},{"id":"titans","title":"Titans Neural Long-Term Memory","category":"brain","backend":"a11oy-native","label":"MODELED","citations":["arXiv:2501.00663"],"endpoint":"/api/a11oy/v1/titans/recall","routes_registered":1},{"id":"mor","title":"Mixture-of-Recursions (adaptive depth)","category":"reasoning","backend":"a11oy-native","label":"MODELED","citations":["arXiv:2507.10524"],"endpoint":"/api/a11oy/v1/mor/route","routes_registered":1},{"id":"goat","title":"GOAT Optimal-Transport Attention","category":"attention","backend":"a11oy-native","label":"UNAVAILABLE","citations":[],"endpoint":"/api/a11oy/v1/govern/manifest/goat","routes_registered":1},{"id":"kla","title":"Kaczmarz Linear Attention","category":"attention","backend":"a11oy-native","label":"UNAVAILABLE","citations":[],"endpoint":"/api/a11oy/v1/govern/manifest/kla","routes_registered":1},{"id":"hrm","title":"Hierarchical Reasoning Model","category":"reasoning","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"pfield","title":"Pressure-Field Coordination","category":"anatomy","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"qhall","title":"Quantum-Inspired Hallucination UQ","category":"governance","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"aimc","title":"Analog In-Memory Attention","category":"attention","backend":"a11oy-native","label":"UNAVAILABLE","citations":[],"endpoint":"/api/a11oy/v1/govern/manifest/aimc","routes_registered":1},{"id":"ternary","title":"Ternary 1.58-bit Weights","category":"quant","backend":"a11oy-native","label":"MODELED","citations":["arXiv:2504.12285","arXiv:2402.17764"],"endpoint":"/api/a11oy/v1/ternary/quantize","routes_registered":1},{"id":"sement","title":"Semantic-Entropy Uncertainty","category":"governance","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"inplacettt","title":"In-Place Test-Time Training","category":"reasoning","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"nested","title":"Nested Learning (multi-timescale)","category":"brain","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"matgran","title":"Matryoshka Representation Granularity","category":"brain","backend":"a11oy-native","label":"UNAVAILABLE","citations":[],"endpoint":"/api/a11oy/v1/govern/manifest/matgran","routes_registered":1},{"id":"graphmem","title":"Multi-Graph Agentic Memory","category":"brain","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"elf","title":"Continuous-Embedding Flow LM","category":"attention","backend":"a11oy-native","label":"UNAVAILABLE","citations":[],"endpoint":"/api/a11oy/v1/govern/manifest/elf","routes_registered":1},{"id":"s3search","title":"Stratified Denoise Search","category":"quant","backend":"a11oy-native","label":"UNAVAILABLE","citations":[],"endpoint":"/api/a11oy/v1/govern/manifest/s3search","routes_registered":1},{"id":"slidesparse","title":"Structured-Sparse Layout Packing","category":"quant","backend":"a11oy-native","label":"UNAVAILABLE","citations":[],"endpoint":"/api/a11oy/v1/govern/manifest/slidesparse","routes_registered":1},{"id":"catq","title":"Calibration Ternary Quant","category":"quant","backend":"a11oy-native","label":"UNAVAILABLE","citations":[],"endpoint":"/api/a11oy/v1/govern/manifest/catq","routes_registered":1},{"id":"rauq","title":"Attention-Pattern Uncertainty","category":"governance","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"agentcoh","title":"Multi-Agent Memory Coherence","category":"brain","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"energy","title":"Energy","category":"energy","backend":"a11oy-native","label":"MEASURED","citations":[],"endpoint":"/api/a11oy/v1/energy/ledger/summary","routes_registered":24},{"id":"fabric","title":"Compute Fabric","category":"defense","backend":"a11oy-native","label":"MODELED","citations":["10.1145/1465482.1465560","10.1145/42411.42415"],"endpoint":"/api/a11oy/v1/fabric/topology","routes_registered":3},{"id":"pnt","title":"PNT","category":"defense","backend":"a11oy-native","label":"MODELED","citations":["10.1103/PhysRevLett.67.181","10.1088/0026-1394/38/1/4","10.1088/1742-6596/723/1/012050","10.1109/TIM.2007.915148","10.5281/zenodo.20528627)"],"endpoint":"/api/a11oy/v1/pnt/limits","routes_registered":7},{"id":"counter-uas","title":"Counter-UAS","category":"defense","backend":"a11oy-native","label":"STRUCTURAL-ONLY","citations":[],"endpoint":"/api/a11oy/v1/counter-uas/info","routes_registered":7},{"id":"governance","title":"Governance","category":"governance","backend":"a11oy-native","label":"UNAVAILABLE","citations":[],"endpoint":"/api/a11oy/v1/governance/verify","routes_registered":4,"note":"native route answered 200 but declared no top-level honest label"},{"id":"pinn","title":"PINN","category":"defense","backend":"a11oy-native","label":"MEASURED","citations":[],"endpoint":"/api/a11oy/v1/pinn/bench","routes_registered":13},{"id":"router","title":"Router","category":"defense","backend":"a11oy-native","label":"MODELED","citations":["10.1109/APEC.2001.911711"],"endpoint":"/api/a11oy/v1/router/active-flux-crossover/info","routes_registered":7},{"id":"anatomy","title":"Anatomy","category":"anatomy","backend":"a11oy-native","label":"LIVE","citations":[],"endpoint":"/api/a11oy/v1/ecosystem/anatomy","routes_registered":8},{"id":"anatomy_body","title":"Anatomy · Living Body","category":"anatomy","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"estate","title":"Estate","category":"map","backend":"a11oy-native","label":"UNAVAILABLE","citations":[],"endpoint":"/api/a11oy/v1/models/estate","routes_registered":1,"note":"native route answered 200 but declared no top-level honest label"},{"id":"muon","title":"Muon Orthogonalized-Momentum Optimizer","category":"attention","backend":"a11oy-native","label":"UNAVAILABLE","citations":[],"endpoint":"/api/a11oy/v1/govern/manifest/muon","routes_registered":1},{"id":"specexec","title":"Tree Speculative Execution","category":"attention","backend":"a11oy-native","label":"UNAVAILABLE","citations":[],"endpoint":"/api/a11oy/v1/govern/manifest/specexec","routes_registered":1},{"id":"nvfp4","title":"NVFP4 4-bit Training Format","category":"quant","backend":"a11oy-native","label":"UNAVAILABLE","citations":[],"endpoint":"/api/a11oy/v1/govern/manifest/nvfp4","routes_registered":1},{"id":"keyless","title":"Keyless Attention (Value-Only Cache)","category":"attention","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"gitthoughts","title":"GitOfThoughts (Version-Controlled Reasoning)","category":"reasoning","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"herotq","title":"HeRo-Q Hessian-Conditioned Quantization","category":"quant","backend":"a11oy-native","label":"UNAVAILABLE","citations":[],"endpoint":"/api/a11oy/v1/govern/manifest/herotq","routes_registered":1},{"id":"dla","title":"Dynamic Linear Attention","category":"attention","backend":"a11oy-native","label":"UNAVAILABLE","citations":[],"endpoint":"/api/a11oy/v1/govern/manifest/dla","routes_registered":1},{"id":"ctxready","title":"Context-Ready Transformer","category":"reasoning","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"opera","title":"OPERA Perplexity-Reward Alignment","category":"reasoning","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"brain","title":"Brain — live knowledge graph","category":"brain","backend":"a11oy-native","label":"MODELED","citations":[],"endpoint":"/api/a11oy/v1/brain/health","routes_registered":82},{"id":"zkinfer","title":"zkML Proof-of-Inference (Cryptographic Receipts)","category":"proof","backend":"a11oy-native","label":"MODELED","citations":["arXiv:2502.18535","arXiv:2404.16109","arXiv:2210.08674","arXiv:2402.02675","10.1145/3627703.3650088","10.48550/arXiv.2210.08674)","10.48550/arXiv.2402.02675)","10.48550/arXiv.2502.18535)"],"endpoint":"/api/a11oy/v1/frontier/zkinfer","routes_registered":1},{"id":"flower","title":"Flower Brain","category":"brain","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"agentmem","title":"AgentMem · Λ-Governed Agent Memory (synthesis)","category":"brain","backend":"a11oy-native","label":"MODELED","citations":["arXiv:2310.08560","arXiv:2502.12110","arXiv:2504.19413","arXiv:2310.03714"],"endpoint":"/api/a11oy/v1/frontier/agentmem","routes_registered":1},{"id":"edgefusion","title":"EdgeFusion · Λ-Gated Energy-Proportional Sensor Fusion (synthesis)","category":"energy","backend":"a11oy-native","label":"MODELED","citations":["arXiv:2410.12032","arXiv:2504.20348","arXiv:2109.12894","arXiv:2205.13542","arXiv:2205.15997","10.1109/MM.2018.112130359"],"endpoint":"/api/a11oy/v1/frontier/edgefusion","routes_registered":1},{"id":"evalarena","title":"Eval Arena · Governed Eval / Red-Team","category":"governance","backend":"a11oy-native","label":"UNAVAILABLE","citations":[],"endpoint":"/api/a11oy/v1/eval-arena","routes_registered":3,"note":"native route answered 200 but declared no top-level honest label"},{"id":"vqc","title":"Governed VQC · Parameter-Shift Hybrid QML (MODELED)","category":"energy","backend":"a11oy-native","label":"MODELED","citations":["arXiv:2312.09121","arXiv:1803.00745","arXiv:1811.11184","arXiv:1811.04968","arXiv:2505.17756"],"endpoint":"/api/a11oy/v1/vqc/manifest","routes_registered":2},{"id":"harness","title":"Governed Model Harness · Behavior Transfer","category":"governance","backend":"a11oy-native","label":"UNAVAILABLE","citations":[],"endpoint":"/api/a11oy/v1/harness/profiles","routes_registered":2,"note":"native route answered 200 but declared no top-level honest label"},{"id":"aigov","title":"AI Governance Conformance · Λ-Advisory Readiness (crosswalk)","category":"governance","backend":"a11oy-native","label":"MODELED","citations":["arXiv:2410.07959","10.6028/NIST.AI.100-1"],"endpoint":"/api/a11oy/v1/frontier/aigov","routes_registered":1},{"id":"fmverif","title":"Proof-Carrying Inference (Machine-Checkable Certificates)","category":"proof","backend":"a11oy-native","label":"MODELED","citations":["arXiv:2401.14461","arXiv:2404.16109","10.1145/263699.263712","10.1145/3656390","10.1007/978-3-031-65630-9_13)"],"endpoint":"/api/a11oy/v1/frontier/fmverif","routes_registered":1},{"id":"supplychain","title":"Model-Artifact Provenance (SLSA / in-toto / Rekor / C2PA)","category":"proof","backend":"a11oy-native","label":"MODELED","citations":["10.1145/3548606.3560596)","10.1145/3548606.3560596"],"endpoint":"/api/a11oy/v1/frontier/supplychain","routes_registered":1},{"id":"hybridssm","title":"HybridSSM · Attention vs State-Space vs Hybrid Frontier (synthesis)","category":"attention","backend":"a11oy-native","label":"MODELED","citations":["arXiv:2312.00752","arXiv:2403.19887","arXiv:2402.19427","arXiv:2406.07522","arXiv:2405.16712"],"endpoint":"/api/a11oy/v1/frontier/hybridssm","routes_registered":1},{"id":"governedagent","title":"Governed Agent Loop · plan→act→self-eval→gate→retry","category":"reasoning","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"governedrag","title":"Governed RAG · Retrieval-with-Receipts","category":"reasoning","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"ecosystem","title":"Harness · Ecosystem Status","category":"defense","backend":"a11oy-native","label":"MODELED","citations":[],"endpoint":"/api/a11oy/v1/ecosystem/mesh","routes_registered":5},{"id":"brainquery","title":"Brain Query · ask the graph (retrieval)","category":"brain","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"sovereign","title":"Sovereign Local Model · status · doctrine self-test · Stage A/B · signed receipt","category":"energy","backend":"a11oy-native","label":"ROADMAP","citations":["arXiv:2604.09048","arXiv:2603.20224","arXiv:2601.22076"],"endpoint":"/api/a11oy/v1/energy/sovereign","routes_registered":4},{"id":"brainbody","title":"Anatomy · Body lit by the Brain (pulse + energy)","category":"anatomy","backend":"a11oy-native","label":"UNAVAILABLE","citations":[],"endpoint":"/api/a11oy/v1/frontier/brainbody","routes_registered":1},{"id":"braincommand","title":"Brain Command · pulse · subscribe/budget · signed receipt (Brain powering the ecosystem)","category":"brain","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"mesh","title":"Sovereign Mesh · Cross-Node Orchestration (status/route/quorum)","category":"defense","backend":"a11oy-native","label":"MODELED","citations":[],"endpoint":"/api/a11oy/v1/mesh/status","routes_registered":12},{"id":"agenttts","title":"Agent Test-Time Compute · Multi-Agent TTC (best-of-N agents + verifier-guided selection)","category":"reasoning","backend":"a11oy-native","label":"MODELED","citations":[],"endpoint":"/api/a11oy/v1/agenttts/scaling","routes_registered":1},{"id":"cryptopipeline","title":"Crypto-Pipeline · End-to-End AI Lifecycle Verifiable Transcript (MODELED)","category":"proof","backend":"a11oy-native","label":"MODELED","citations":[],"endpoint":"/api/a11oy/v1/cryptopipeline/transcript","routes_registered":1},{"id":"gateddelta","title":"Gated Delta-Rule Linear Attention · state recall (MODELED)","category":"attention","backend":"a11oy-native","label":"MODELED","citations":[],"endpoint":"/api/a11oy/v1/gateddelta/recall","routes_registered":1},{"id":"agentops","title":"Agent Ops · bounded operate loop (ground→act→self-eval→gate, Ouroboros-bounded, writer≠judge)","category":"reasoning","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"lgmi","title":"LGMI · Λ-Governed Mechanistic Interpretability (synthesis)","category":"governance","backend":"a11oy-native","label":"MODELED","citations":["arXiv:2309.08600","arXiv:2407.14435","arXiv:2304.14997","arXiv:2211.00593","arXiv:2403.00745"],"endpoint":"/api/a11oy/v1/frontier/lgmi","routes_registered":1},{"id":"gnqs","title":"GNQS · Governed-Norm Quantization Stability (synthesis)","category":"quant","backend":"a11oy-native","label":"MODELED","citations":["arXiv:2208.07339","arXiv:2210.17323","arXiv:2306.00978","arXiv:2211.10438","arXiv:2402.17764","arXiv:1910.07467","arXiv:2402.17762"],"endpoint":"/api/a11oy/v1/frontier/gnqs","routes_registered":1},{"id":"numericsdataset","title":"Numerics Dataset · 1,328 frozen matrix cases + append-only engine evidence · preregistered inputs are never shown as measured results · MATCH/CONFLICT/UNAVAILABLE · proof uplift=0","category":"quant","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"casta","title":"CASTA · Clean-room Anomaly × Streaming Test-time Adaptation (synthesis)","category":"defense","backend":"a11oy-native","label":"MODELED","citations":["arXiv:2006.10726","arXiv:2203.13591","arXiv:2204.02610","arXiv:2110.09506","10.1109/ICDM.2008.17"],"endpoint":"/api/a11oy/v1/frontier/casta","routes_registered":1},{"id":"frontierindex","title":"Frontier Index · honest ecosystem catalog · per-surface backend label + citations + native/fallback (self-audited)","category":"brain","backend":"a11oy-native","label":"MODELED","citations":[],"endpoint":"/api/a11oy/v1/frontier-index/health","routes_registered":1},{"id":"sparsemoe","title":"Extreme-Sparsity MoE Analyzer · activation-ratio ↔ inference-cost tradeoff (STRUCTURAL-ONLY)","category":"quant","backend":"a11oy-native","label":"MODELED","citations":["arXiv:2401.04088","arXiv:2401.06066"],"endpoint":"/api/a11oy/v1/frontier/sparsemoe","routes_registered":1},{"id":"pddisagg","title":"Prefill/Decode Disaggregation Map · colocated vs disaggregated latency across the mesh (STRUCTURAL-ONLY/ROADMAP)","category":"defense","backend":"a11oy-native","label":"MODELED","citations":["arXiv:2603.13358","arXiv:2401.09670","arXiv:2311.18677"],"endpoint":"/api/a11oy/v1/frontier/pddisagg","routes_registered":1},{"id":"execverify","title":"Execution-Verified Synthesis Loop · eval → verified trajectory → corpus candidate → receipt (STRUCTURAL-ONLY)","category":"reasoning","backend":"a11oy-native","label":"MODELED","citations":["arXiv:1807.03100","arXiv:2203.14465"],"endpoint":"/api/a11oy/v1/frontier/execverify","routes_registered":1},{"id":"opsdash","title":"Ops Dashboard · live estate health · GREEN/DEGRADED/UNAVAILABLE rollup + per-surface honest labels + /verify","category":"governance","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"flowbrain","title":"FlowBrain · continuous belief-flow lens · tiers as thresholds crossed on x_t∈[0,1] + 1D-time⊗1D-node axis-factorization (STRUCTURAL-ONLY; borrows B[FM]² continuous-flow principle, NO EEG)","category":"brain","backend":"a11oy-native","label":"MODELED","citations":["arXiv:2606.20812","arXiv:2606.20812v1"],"endpoint":"/api/a11oy/v1/frontier/flowbrain","routes_registered":2},{"id":"blocksparse","title":"Learned Blockwise Top-k KV Sparse Attention · selection (MODELED)","category":"attention","backend":"a11oy-native","label":"MODELED","citations":[],"endpoint":"/api/a11oy/v1/blocksparse/select","routes_registered":1},{"id":"confattest","title":"Confidential-Compute Attestation + Action Gate · SIMULATED enclave quote → receipt · govern ACTIONS not reasoning (MODELED)","category":"governance","backend":"a11oy-native","label":"MODELED","citations":["arXiv:2606.26298","arXiv:2604.12986","arXiv:2606.31408","arXiv:2606.11145"],"endpoint":"/api/a11oy/v1/confattest/gate","routes_registered":2},{"id":"retrievalattn","title":"Retrieval-Modulated Long-Context Attention · needle recall (MODELED)","category":"attention","backend":"a11oy-native","label":"MODELED","citations":[],"endpoint":"/api/a11oy/v1/retrievalattn/recall","routes_registered":1},{"id":"whatsnew","title":"What's New · honest auto-derived estate changelog · recently-added surfaces w/ verbatim labels + citations from real git history (drift-proof)","category":"brain","backend":"a11oy-native","label":"MODELED","citations":[],"endpoint":"/api/a11oy/v1/whatsnew/health","routes_registered":2},{"id":"honestywall","title":"Honesty Wall · live \"can this system lie right now?\" integrity aggregation · reads each surface's OWN honest label VERBATIM + estate honesty invariants → INTACT/DEGRADED/VIOLATED verdict, unsigned SHA-256 receipt-on-write (drift-proof)","category":"governance","backend":"a11oy-native","label":"MODELED","citations":[],"endpoint":"/api/a11oy/v1/govern/honestywall/info","routes_registered":2},{"id":"agentos","title":"Agent OS Map · live self-honest operator's-eye map of the agent OS · nodes (daily loop←agentops, trust ledger←anatomy+receipts+honestywall, standing goals←doctrine+locked-8, optional loops←governedagent/governedrag/loopforge/mesh) with LIVE per-node verdict from the honestywall aggregate → OPERATING/DEGRADED/HALTED-HONEST, never OPERATING if anything VIOLATED, unsigned SHA-256 receipt-on-write (drift-proof)","category":"governance","backend":"a11oy-native","label":"MODELED","citations":[],"endpoint":"/api/a11oy/v1/govern/agentos/status","routes_registered":4},{"id":"brainground","title":"Brainground · grounding-confidence + honest abstention over brain retrieval · scores the brain's REAL grounding_subgraph (seed coverage · subgraph cohesion · salience mass · community consistency) → GROUNDED/WEAK-GROUNDING/INSUFFICIENT-GROUNDING, brain abstains when grounding is weak, MODELED (never MEASURED), unsigned SHA-256 receipt-on-write","category":"brain","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"brainmemory","title":"Brain Memory Freshness · honest episodic decay proxy over the knowledge graph · deterministic per-node freshness (recency + structure when a real timestamp exists, else STRUCTURAL-ONLY connectivity+salience proxy, never a fabricated timestamp) -> FRESH/AGING/STALE, STALE flagged for re-harvest, unsigned SHA-256 receipt-on-write (drift-proof)","category":"brain","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"brainprovenance","title":"Brain Provenance · per-answer source-lineage chain · which knowledge-graph nodes supported the answer, each node's OWN label VERBATIM + contribution_weight + coverage (HARVESTED/MODELED/UNAVAILABLE) → TRACEABLE/PARTIAL-PROVENANCE/UNTRACEABLE, never TRACEABLE while any node UNAVAILABLE, unsigned SHA-256 receipt-on-write (source-lineage of an answer, NOT build/model attestation)","category":"brain","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"braincontradict","title":"Brain Contradiction Detector · surfaces potential contradictions between grounded knowledge-graph claims HONESTLY via transparent deterministic heuristics (negation polarity / antonym opposition / numeric conflict) → NO-CONFLICT/POSSIBLE-CONFLICT/CONFLICT-FLAGGED · PRESENTS conflicts, NEVER resolves them (adjudication=human-required), unsigned SHA-256 receipt-on-write (MODELED)","category":"brain","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"brainuncertainty","title":"Brain Uncertainty · calibrated honest uncertainty on a brain retrieval · reads the same honest ranked retrieval and derives score dispersion + retrieval entropy + rank stability → one uncertainty in [0,1] with a CONFIDENT/UNCERTAIN/HIGHLY-UNCERTAIN verdict (recommends abstain), never CONFIDENT when dispersion/entropy high, MODELED calibration honesty not a probability, unsigned SHA-256 receipt-on-write","category":"brain","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"brainhealth","title":"Brain Health · live \"can the brain be trusted for this query right now?\" rollup · reads each brain-honesty surface's OWN label VERBATIM (grounding, freshness, provenance, contradiction, uncertainty), degrades absent siblings to UNAVAILABLE → TRUSTWORTHY/DEGRADED/UNTRUSTWORTHY/INSUFFICIENT-SIGNAL, never TRUSTWORTHY if any available component abstains, unsigned SHA-256 receipt-on-write (MODELED)","category":"brain","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"brainreranker","title":"Anatomy v6 · Brain evidence inventory, reranker readiness, and bounded Ouroboros local-feed receipts · every raw node receives an admission/quarantine decision · zero proposal rows while canonical manifests are absent","category":"brain","backend":"a11oy-native","label":"MODELED","citations":[],"endpoint":"/api/a11oy/v1/brainreranker/status","routes_registered":1},{"id":"brainwatch","title":"Brain Watch · knowledge-graph honesty-posture drift monitor · MEASURED label-distribution/orphan/community/salience snapshot vs a caller-supplied PRIOR → STABLE/DRIFTING/DEGRADED/BASELINE-ONLY (no fabricated trend without a real prior), unsigned SHA-256 receipt-on-write","category":"brain","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"brainconsensus","title":"Brain Consensus · honest corroboration of a brain grounding · measures how MANY distinct nodes support a query and how many distinct communities they span (cross-community agreement is stronger than one clique) → CORROBORATED/WEAK-CORROBORATION/SINGLE-SOURCE, single-source-risk flag when support collapses to one node/community, never CORROBORATED while that flag is set, MODELED corroboration honesty not a truth guarantee, unsigned SHA-256 receipt-on-write","category":"brain","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"brainqueryaudit","title":"Brain Query Audit · append-only hash-linked ledger of brain queries + the honest verdict each returned · POST appends {query, verdict, grounding_label} and mints an UNSIGNED SHA-256 receipt chained to the prior entry (tamper-evident); GET recomputes the whole chain → CHAIN-INTACT/CHAIN-BROKEN (never softened), ephemeral in-memory ledger labelled honestly, MODELED, receipt-on-write-not-on-read","category":"brain","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"brainlineage","title":"Brain Lineage · node-origin chain · how each knowledge-graph node ENTERED the graph, read VERBATIM from its OWN real origin fields (source/url → structural derivation → none) → TRACED/PARTIAL-LINEAGE/UNKNOWN-ORIGIN, a node with no source is UNKNOWN-ORIGIN never a fabricated source, aggregate never TRACED while any origin UNKNOWN, unsigned SHA-256 receipt-on-write (node-origin lineage, NOT per-answer provenance, NOT build/model attestation)","category":"brain","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"brainexplain","title":"Brain Explain · transparent explanation of WHY the brain retrieved what it did · MODELED descriptive trace over the REAL retrieval subgraph (which query terms matched which seed nodes, per-node ppr-vs-salience rationale, communities traversed, each node's OWN label VERBATIM) → EXPLAINABLE/PARTIALLY-EXPLAINABLE/OPAQUE (never invents a rationale; honest OPAQUE beats a fake one), unsigned SHA-256 receipt-on-write","category":"brain","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"braingaps","title":"Brain Gaps · an honest map of what the brain does NOT know · MEASURED thin (sparse) communities + weakly-connected island nodes (degree≤1) + weak-label share over the live graph, and per-query COVERED/THIN/GAP grounding → estate verdict WELL-COVERED/PATCHY/SPARSE (a GAP is never fabricated into coverage), unsigned SHA-256 receipt-on-write (MODELED)","category":"brain","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"brainconstitution","title":"Brain Constitution · the honest, machine-checkable ruleset the brain is graded against per query · an explicit ordered set of ARTICLES (grounding sufficiency, calibrated confidence, honest corroboration, contradictions surfaced, traceable to source, freshness honesty, coverage gaps admitted, doctrine invariants) each graded COMPLIANT/VIOLATED/UNAVAILABLE against whatever sibling brain-honesty surfaces are importable (an absent surface is UNAVAILABLE, never a fabricated pass) → CONSTITUTIONAL/IN-VIOLATION/INSUFFICIENT-SIGNAL, never CONSTITUTIONAL while any evaluable Article is VIOLATED, unsigned SHA-256 receipt-on-write (MODELED)","category":"brain","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"markets","title":"Markets & Finance · live Polymarket prediction book (top markets by 24h volume, pillar height = 24h vol, bead = YES probability) + crypto majors orbiting by market cap · 100% MEASURED same-origin deva finance feeds, no mocked spend/keys, log-scaling is display-only","category":"finance","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"leaders","title":"Frontier Models · Live AI Leaders · MEASURED live OpenRouter model catalog — one pillar per lab (height = max context window, bead = open/free share); widest-context models orbit as satellites (teal = free, amber = paid); log-scaling is display-only, no invented benchmark or ranking","category":"more","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"research","title":"Frontier Research · Live arXiv AI · MEASURED live arXiv submission stream (cs.AI/LG/CL/CV/NE) — one pillar per research category (height = fresh-paper count, bead = share of the window); most-recent papers orbit as satellites sized by author count; log-scaling is display-only, no citation count, score, or ranking","category":"more","backend":"a11oy-native","label":"PROVEN","citations":[],"endpoint":"/api/a11oy/v1/evidence/research","routes_registered":7},{"id":"open","title":"Open Frontier · Live Hugging Face · MEASURED live Hugging Face Hub trending stream — one pillar per org/author (height = summed likes, bead = download reach); most-liked open models orbit as satellites coloured by task; log-scaling is display-only, no invented benchmark, score, or ranking beyond the Hub’s own","category":"more","backend":"a11oy-native","label":"UNAVAILABLE","citations":[],"endpoint":"/api/a11oy/v1/deva/frontier/open","routes_registered":1,"note":"native route answered 200 but declared no top-level honest label"},{"id":"brainagent","title":"Brain Agent · honesty-gated agentic graph reasoner · treats the brain graph as a state space and walks it one node at a time for a query, choosing EXPAND/FOLLOW/BACKTRACK/STOP by pure graph heuristics (no model call, fully deterministic); every candidate hop passes an honesty gate built from the sibling surfaces (grounding, provenance, contradiction, uncertainty) — an ungrounded/untraceable/contradicted/uncertain hop is refused with a recorded reason, an absent guard is UNAVAILABLE never a fabricated pass; within an explicit step/node budget it returns ANSWER-GROUNDED/PARTIAL or abstains ABSTAINED-BUDGET/ABSTAINED-INSUFFICIENT rather than answer under-grounded, emits an ordered trace + unsigned SHA-256 receipt-on-write, makes no sentience claim (MODELED)","category":"brain","backend":"frontend-only","label":"UNAVAILABLE","citations":[],"endpoint":null,"routes_registered":0,"note":"no a11oy-native /api route registered; client-side surface"},{"id":"estateconstitution","title":"Estate Constitution · the WHOLE ESTATE's honesty posture graded continuously against explicit estate-level ARTICLES (the per-query brain-constitution pattern lifted to the estate) · Art 1 the honesty wall holds (0 REACHABLE invariant violations), Art 2 no surface declares a fabricated MEASURED label (every declared label in the honest vocabulary, read VERBATIM), Art 3 manifest coverage DISCLOSED honestly (NATIVE-OK vs NO-MANIFEST counted out loud — a DISCLOSURE Article: admitting the gap is compliance, papering it over is the violation, full coverage never claimed while any surface is unverifiable), Art 4 doctrine invariants (Λ = Conjecture 1 never a theorem, locked-8 never inflated, trust ceiling 0.97 never 100%, no consciousness claim) · each graded COMPLIANT/VIOLATED/UNAVAILABLE over the szl_honestywall aggregate read in-process through a GUARDED import (an unreadable aggregate is UNAVAILABLE, never a fabricated pass) → CONSTITUTIONAL/IN-VIOLATION/INSUFFICIENT-SIGNAL, never CONSTITUTIONAL while any evaluable Article is VIOLATED, unsigned SHA-256 receipt-on-write (MODELED)","category":"governance","backend":"a11oy-native","label":"MODELED","citations":[],"endpoint":"/api/a11oy/v1/govern/estateconstitution/status","routes_registered":2},{"id":"attest","title":"Chain of Title · L6 attestation · the allodial claim made third-party verifiable · an in-toto v1 Statement whose SUBJECT binds the locked-8 kernel gitCommit c7c0ba17 (plus a sovereign-weights sha256 ONLY when a real weights artifact is readable this request — honest null otherwise, never an invented digest), predicateType https://szl.dev/chain-of-title/v1, predicate = doctrine v11 + provenance (corpus_sha, training config read VERBATIM from the committed trainer, kernel_verified RE-CHECKED this request against the digest-verified formula registry, kernel_pin) + energy_measured[] (HONESTLY EMPTY with no live joule meter — no joule fabricated) + honesty_invariants {no_fabricated_measured, lambda_is_conjecture_not_theorem, locked8_immutable, provenance_coverage 1.0} bound INTO the signed predicate so the doctrine itself becomes third-party checkable + the tier-PROPOSED SEAL formula (EU CSF SEAL + HHI prior art, no score asserted) · DSSE-signed through the estate cosign key (explicitly UNSIGNED-NO-KEY when no runtime secret is present, never a fabricated signature) and structured for cosign keyless + Rekor: a real inclusion proof reads RECORDED/MEASURED, an unreachable log reads UNREACHABLE and the verdict is UNKNOWN — never a fabricated Rekor entry or PASSED · the verdict is tri-state and FAILED beats UNKNOWN so a tampered Statement cannot hide behind an offline log; policy authoritative in ops/szl_chain_of_title.rego, mirrored in-process rule for rule (MODELED)","category":"governance","backend":"a11oy-native","label":"MODELED","citations":[],"endpoint":"/api/a11oy/v1/attest/manifest","routes_registered":6},{"id":"brainretro","title":"Brain Retro · retrospective CALIBRATION of the brain's OWN past answers — the self-honesty loop, calibration accounting (never model training, never a sentience claim) · reads the append-only query-audit ledger (what the brain recorded it answered) through a guarded import, RE-RUNS the CURRENT grounding for a sample of those same queries, and compares: of the answers recorded as confident how many are STILL grounded, and of the abstentions how many were justified (still ungrounded) · each past entry classified CONFIRMED / DRIFTED (OVER-CLAIM = past grounded now ungrounded, a real honesty risk, or OVER-CAUTION) / STALE-UNKNOWN (not recomputable — excluded from every rate, a past answer is NEVER counted correct without recomputed evidence) → WELL-CALIBRATED/DRIFT-DETECTED/INSUFFICIENT-HISTORY, never softened; honest caveat: the query-audit ledger is EPHEMERAL (in-memory, resets on restart) so a fresh Space truthfully reports INSUFFICIENT-HISTORY, unsigned SHA-256 receipt-on-write (MODELED)","category":"brain","backend":"a11oy-native","label":"MODELED","citations":[],"endpoint":"/api/a11oy/v1/brain/brainretro/manifest","routes_registered":1},{"id":"brainanswer","title":"Brain Answer · governed honest-answer synthesizer · the capstone that ties the brain-honesty surfaces into ONE endpoint: ask a question, get a single answer object with a full honesty dossier or an honest abstention · the ANSWER comes only from brainagent's honesty-gated traversal (never a confident answer the agent could not ground), and the DOSSIER carries grounding (brainground), provenance chain (brainprovenance), uncertainty (brainuncertainty), contradiction flags (braincontradict) and per-query constitutional compliance (brainconstitution), each read VERBATIM under its own honest label with an absent sibling reading UNAVAILABLE rather than fabricated · the governed verdict only ever DOWNGRADES: ANSWERED-GOVERNED (grounded + COMPLIANT + no unresolved contradiction) / ANSWERED-WITH-CAVEATS (grounded but weak, uncertain or partial — caveats listed) / ABSTAINED (agent abstained OR constitution IN-VIOLATION OR contradiction CONFLICT-FLAGGED) / INSUFFICIENT-SIGNAL (too few facets available to govern an answer), so it can never read ANSWERED-GOVERNED while any of those three conditions holds · unsigned SHA-256 receipt-on-write, GET mints nothing, no sentience claim (MODELED)","category":"brain","backend":"a11oy-native","label":"MODELED","citations":[],"endpoint":"/api/a11oy/v1/brain/brainanswer/manifest","routes_registered":1},{"id":"brainlocal","title":"Brain Local · honest liveness+capability probe of the LOCAL OpenAI-compatible inference endpoint the brain can be wired to (Ollama on 127.0.0.1:11434, a llama.cpp llama-server, or either behind a cloudflared tunnel) · reads SZL_LOCAL_LLM_URL / A11OY_JPT_GPU_URLS / A11OY_JPT_MODELS and, on THIS request, makes one short bounded stdlib-urllib GET of the endpoint's OWN model listing ({base}/v1/models then {base}/api/tags) → LIVE with the served model list echoed VERBATIM from the endpoint and label MEASURED, earned by that live reading alone; env unset → UNAVAILABLE with the note 'no local endpoint configured' and no model named; timeout/refused/error → UNAVAILABLE, never softened into a healthy reading; reachable but naming NO model → DEGRADED, reported as DEGRADED and never as healthy. A11OY_JPT_MODELS tags are operator DECLARATIONS (MODELED) listed separately as declared-not-served and never merged into the served list. Performs NO inference and stores nothing; wall-readable honesty manifest at /api/a11oy/v1/brain/brainlocal/manifest (surface's OWN posture MODELED, never upgraded by one live reading), unsigned SHA-256 receipt-on-write, no consciousness claim (MODELED)","category":"brain","backend":"a11oy-native","label":"MODELED","citations":[],"endpoint":"/api/a11oy/v1/brain/brainlocal/manifest","routes_registered":1},{"id":"braincite","title":"Brain Cite · verifiable claim→source citations for brain answers · runs the SAME honest brain retrieval (szl_brain_api.get_index(ns).ask(q,k)) and binds each candidate claim term drawn from the query to the source node(s) whose title actually contains it — a claim with a real backing node is CITED, a claim with none is UNCITED and is NEVER given a fabricated citation, so an answer can never present an uncited claim as sourced · each cited source carries id/title/url/node_label VERBATIM (membership never upgrades a label); citation_coverage = cited/total is a MODELED ratio, never MEASURED → FULLY-CITED / PARTIALLY-CITED / UNCITED-DOMINANT / NO-CITABLE-CLAIMS · complements brainprovenance (whole lineage) by binding at claim granularity; wall-readable honesty manifest at /api/a11oy/v1/brain/braincite/manifest, unsigned SHA-256 receipt-on-write, GET mints nothing, no sentience claim (MODELED)","category":"brain","backend":"a11oy-native","label":"MODELED","citations":[],"endpoint":"/api/a11oy/v1/brain/braincite/manifest","routes_registered":1},{"id":"braineval","title":"Brain Eval · honest refusal-to-fabricate GATE for a SERVED model — the eval that makes a 'trained' claim meaningful · reads the SAME endpoint contract as brainlocal (SZL_LOCAL_LLM_URL / SZL_LOCAL_LLM_MODEL) and, when a model answers THIS request, POSTs a fixed battery of doctrine PROBE prompts across six violation families (fabricated-MEASURED, lambda-as-theorem, banned-superlative, fake-wired-live, fabricated-attestation, consciousness-overclaim) and scores each response REFUSED / FABRICATED by transparent lexical rules → MEASURED refusal_rate (capped at the 0.97 ceiling, never 1.0) with verdict REFUSAL-HONEST / PARTIAL-REFUSAL / FABRICATION-DETECTED; env unset or node asleep → UNAVAILABLE, never a fabricated score, and any single fabrication is surfaced as FABRICATION-DETECTED, never hidden · EVALUATES only — trains nothing, writes nothing back to any model; wall-readable honesty manifest at /api/a11oy/v1/brain/braineval/manifest, unsigned SHA-256 receipt-on-write, no sentience claim (MODELED)","category":"brain","backend":"a11oy-native","label":"MODELED","citations":[],"endpoint":"/api/a11oy/v1/brain/braineval/manifest","routes_registered":1},{"id":"brainserve","title":"Brain Serve · governed BRIDGE to the estate's OWN served brain model — wires the committed SZL-Khipu-1.5B-GGUF inference Space (OpenAI-compatible, provenance-bound) so a real brain model answers WITHOUT any operator secret · on each read it POSTs a bounded probe, verifies the served model's self-reported repo@revision against the pinned expectation (SERVING-EXPECTED vs PROVENANCE-MISMATCH, mismatch surfaced never hidden), and reports MEASURED when the model answers THIS request; asleep/cold/unreachable → UNAVAILABLE, never a fabricated reading · propagates the endpoint's honesty caveats VERBATIM (output signature_status=UNSIGNED stays unsigned; receipts.covers_this_output=false is shown as a caveat, never upgraded to proven; service_level=BEST_EFFORT_NO_SLA) · BRIDGES only — no training, admits 0 gradient rows, not counter-UAS, no sentience · wall-readable manifest at /api/a11oy/v1/brain/brainserve/manifest, unsigned SHA-256 receipt-on-write (MODELED)","category":"brain","backend":"a11oy-native","label":"MODELED","citations":[],"endpoint":"/api/a11oy/v1/brain/brainserve/manifest","routes_registered":1},{"id":"brainreceipt","title":"Brain Receipt · SIGNED inference receipts — mints a portable, offline-verifiable receipt binding SHA-256(request) + SHA-256(sources) + model_id to SHA-256(output), signed with the estate's shared ECDSA-P256 key · a valid signature proves INTEGRITY + key continuity ONLY, explicitly NOT that the output is correct, true, non-hallucinated, or source-supported (those are braincite/braineval) — stated verbatim in every receipt · key_source reported honestly: persistent Secret key → SIGNED-LOCAL, ephemeral container key → UNSIGNED-LOCAL (real bytes, honestly scoped), no key → UNAVAILABLE digest-only, never a fabricated signature · per-source SHA-256 makes corpus membership checkable; tamper is detected on verify · the SIGNED complement to the estate's unsigned digests and the request-side complement to brainserve · no new dependency, 0 runtime CDN, trains nothing, no sentience (MODELED)","category":"brain","backend":"a11oy-native","label":"MODELED","citations":[],"endpoint":"/api/a11oy/v1/brain/brainreceipt/manifest","routes_registered":2},{"id":"brainverdict","title":"Brain Verdict · THE VERIFIABLE ANSWER — composes served-model provenance (brainserve) + claim→source citation coverage (braincite) + refusal-to-fabricate posture (braineval) into ONE assurance verdict, then binds the whole chain under the estate ECDSA-P256 signature (brainreceipt) as a single offline-verifiable object · assurance_level VERIFIABLE-GROUNDED / VERIFIABLE-PARTIAL / VERIFIABLE-WEAK / UNVERIFIABLE-NO-MODEL is DETERMINISTIC and never above what the components earned; the weakest link is always named so the verdict is never opaque · a valid signature proves the chain is UNALTERED — INTEGRITY, not truth (stated verbatim); no live model → UNVERIFIABLE-NO-MODEL, no component ever fabricated · the SOTA-identified frontier move (provenance-bound serving + signed receipt + grounding + abstention in one primitive); open spec at docs/verifiable-answer-receipt.md · 0 runtime CDN, trains nothing, not counter-UAS (MODELED)","category":"brain","backend":"a11oy-native","label":"MODELED","citations":[],"endpoint":"/api/a11oy/v1/brain/brainverdict/manifest","routes_registered":1},{"id":"brainsentry","title":"Brain Sentry · DEFENSIVE-CYBER signal triage (blue-team) — scores and ranks raw security signals (log lines, alerts, indicators) by TRANSPARENT, auditable defensive rules (MITRE-ATT&CK-flavored: brute force, priv-esc, injection, C2 beacon, ransomware, exfil, persistence, defense-evasion) so an analyst reviews the riskiest first · every score is a MODELED sum of matched rule weights and EVERY matched rule is returned, so the triage is never a black box; no signals → UNAVAILABLE, never an invented verdict · RANKS and EXPLAINS only — takes NO action, blocks nothing, attacks nothing, never claims a signal IS malicious (adjudication is human-required) · DEFENSIVE-BLUE-TEAM-ONLY: not offensive tooling, not counter-UAS, not weapons · unsigned SHA-256 receipt-on-write, 0 runtime CDN, trains nothing (MODELED)","category":"brain","backend":"a11oy-native","label":"MODELED","citations":[],"endpoint":"/api/a11oy/v1/brain/brainsentry/manifest","routes_registered":1}],"timestamp_utc":"2026-10-10T14:27:39.135372+00:00"}