feat(devices): multi-device task claiming — only the winning device executes a task

chore: bump device-reported version to 1.5.0

feat(brain): all-science reasoning upgrade — persistent memory injection, few-shot exemplars, actor-critic verification, uncertainty rule

docs: enterprise documentation suite — compliance mappings (CRA, ISO 27001, IEC 62443), threat model, data flow, audit & deployment guides, SBOM, use cases, security.txt

feat(brain): auto smart/cheap mode — per-task brain profiles with provider routing (simple/standard/complex heuristics, quality-first vs cost-first key ordering)

fix(parser): lenient answer salvage — no raw JSON leaks when the brain breaks JSON with unescaped quotes (e.g. German „..." quotes); 3 new tests

feat(reasoning): deep reasoning engine — plan, act, reflect, verify

feat(trace): deep insight — live run traces with per-step usage, tokens, timing and cost

feat(tools): persistent memory + web research tools for the agent

fix(cloud): WAF-safe transport — base64 bodies for agent API payloads

feat: enterprise capability expansion — app control, browser control, background execution

feat(agent): self-healing loop — auto-debug, auto-reasoning, iteration

fix(agent): robust tool-call parsing — prose, fences and nested braces

polish: community touchpoints — star/share links, CLI + installer CTAs

fix(cli): login crash on failed verify — readline used after close

feat(agent): agentic execution loop — the agent that acts

fix(installer): point dashboard URL at /dashboard

feat: desktop app packaging + engine/protocol/gateway packages, HTTP metrics pipeline