fix(honesty): warhacker /index locked_proven = exactly 5; Wave8 demo formulas split to experimental_ci_green

rename: AMARU->YACHAY, SENTRA->CHAPAQ (user-visible) + SLSA reconcile

Warhacker: expand demo tab to full 5x5 = 25 demos (per-problem demos, real tamper, honest formula panels, sovereign canvas globe fallback)

fix(a11oy.code): scrub deep-reasoning co-pilot codename from /v1/code/chat-deep response payload

fix(docker): COPY a11oy_code.py into image so the /api/a11oy/v1/code router surface registers

fix(a11oy.code): register /api/a11oy/v1/code/{tiers,health,index,roster,route,auto,complete} before catch-all; live HF-router inference with honest deterministic fallback; rename chat-with-rosie -> chat-deep (scrub codename)

a11oy: framing pass v2 — flex-column body lock + kbformulas/mobile overflow fix

a11oy: framing pass v2 — flex-column body lock + kbformulas/mobile overflow fix

a11oy: live-data layer + a11oy Code live HF inference + framing/mobile pass

feat: BATCH-3 final tab uniqueness (15 distinct viz) + 2 new Warhacker demos (ai-sbom, agentic-provenance)

feat: BATCH-3 final tab uniqueness (15 distinct viz) + 2 new Warhacker demos (ai-sbom, agentic-provenance)

fix(api): register batch-2 /api/a11oy/v1/sec/* BEFORE the Node proxy catch-all

feat(console): BATCH-2 tab uniqueness (15 distinct real-data viz) + sec data module

feat(console): batch-1 tab uniqueness (distinct real-data viz per tab) + sovereign fonts

feat(console): batch-1 tab uniqueness (distinct real-data viz per tab) + sovereign fonts

feat(console): batch-1 tab uniqueness (distinct real-data viz per tab) + sovereign fonts

feat(console): batch-1 tab uniqueness (distinct real-data viz per tab) + sovereign fonts

chore: serve.py vendor allowlist + Dockerfile per-file COPY + NOTICE attributions

feat(console): batch-1 tab uniqueness (distinct real-data viz per tab) + sovereign fonts

chore: serve.py vendor allowlist + Dockerfile per-file COPY + NOTICE attributions

fix(fonts): self-host @font-face in SPA shells (remove Google Fonts CDN)

feat(vendor): add dagre+cytoscape-dagre+d3-sankey (MIT/ISC) for batch-1 viz

feat(vendor): add ngraph.graph/path/forcelayout+panzoom (BSD-3/MIT, anvaka)

fix(warhacker): make shallow launch endpoint genuinely mode-aware + real per-run receipts

a11oy: wh-demo run handler honors ?mode= query param (robustness)

a11oy: Warhacker EXHAUSTIVE demos (5 full step-by-step demos)

yuyay licenseHygiene: word-boundary match (disclosed != closed)

yuyay heuristic fix: 'never 100%' is honest hedge (not overclaim) + honest preset wording

deep-upgrade: vendor fonts (zero CDN), Prometheus proxy + YUYAY 13-axis conjunctive gate endpoint, Organism anatomy tab + Warhacker play-by-play (step timeline + boolean cascade + tamper-fail + hollow-pass guard)

add capability_map (theorem->mechanism wiring) to a11oy proof_summary

COPY szl_formula_wiring.py into a11oy image (per-file)

register formula-wiring surface on a11oy app (additive, pre-catchall)

add szl_formula_wiring (shared): live theorem mechanisms + endpoints

wire ~80 theorems into governed loop (a11oy): formula_proof per run

FIX: honest proven count + Live System Map redesign (service-health dashboard)

NOTICE: attribute vendored MIT/Apache viz libs (3d-force-graph/globe.gl/force-graph vasturiano MIT, ECharts/echarts-gl Apache-2.0, Cytoscape MIT, d3 ISC) for frontier tabs

a11oy frontier tabs: +7 live 3D tabs (Governance Atlas, AI Oversight, Model Atlas, MELT Observability, Threat Graph, Deploy Posture, Warhacker Proofs) — additive, reads existing live APIs, vendored MIT viz

a11oy frontier tabs: +7 live 3D tabs (Governance Atlas, AI Oversight, Model Atlas, MELT Observability, Threat Graph, Deploy Posture, Warhacker Proofs) — additive, reads existing live APIs, vendored MIT viz

a11oy Code: governed agentic coder tab (chat/code/research) + per-run-genesis tamper-evidence fix

Forge OPEN-WEIGHT alloy model layer into a11oy (DeepSeek CODE_PRIMARY; C20/W7-5 router; W5-3/W7-4 conformal; C10-C12 consensus) — UNIFIED into LLM registry

Vendor viz libs no-CDN (air-gap) + bind graph-substrate proofs to console graphs

console: agentic Exposure Priority panel on CVE Watch (leader-grade, additive)

console: agentic Exposure Priority panel on CVE Watch (leader-grade, additive)

fix(loop): relocate uvicorn.run entry-point to EOF so the governed 6-receipt agent loop + warhacker tabs register (were dead code after a mid-file blocking uvicorn.run); ship proven 6-receipt szl_agentic_loop (retrieve->quarantine->tool->policy->kernel->emit, P3 non-interference). Surgical: only serve.py entry-point moved + module updated. Signed-off-by: Stephen P. Lutar Jr. <stephenlutar2@gmail.com>

a11oy: 5 real mission surfaces (AI Oversight / Deploy Posture / Mission Health / Trajectory Picture / Edge Run)

fix(loop): move GOVERNED AGENT LOOP block BEFORE `if __name__=='__main__'` uvicorn.run() — it was dead code after the blocking call (never executed under `python serve.py`), so /mcp/ + /api/<ns>/v1/agent/* never registered. Now registers at boot.

diag: add /api/a11oy/v1/agent/_diag (inserted at 0) to surface loop-registration result/traceback at runtime

instill: knowledge.json v6.0.0 proof_summary + anatomy, conformal Trust Score, formulas tab

fix(docker): drop COPY of 3 files absent from Space (szl_llm_registry.py, szl_elite_console.py, web/elite_console.html) that caused BUILD_ERROR; keep only the new COPY szl_agentic_loop.py. Rebased on last-good Dockerfile.

feat(loop): operational RAG->tool-call->policy/trust gate->signed-receipt loop + canonical live MCP at /mcp/ + /ask-and-act consumer UI; de-advertise retired hatun-mcp Space on /hatun-mcp page