{"kind":"uaix.roadmap.v1","standard_family":"UAI","current_release":"UAI-1 \/ v1.0","uai_version":"1.0","release_date":"2026-05-31","description":"Forward roadmap for the public UAI-1 \/ v1.0 launch surface, today-focus orientation, Agent Communication Operating Model, roadmap documentation ledger, interoperability evidence, bridge evidence examples, compact-transfer discipline, trust metadata expansion, conformance fixture evidence, conformance maturity, AI Memory starter and wizard outputs, Hierarchical Goal Memory guide\/export support, universal required memory-maintenance files, active instruction anchors, project-context handoff, active file intake, agentic-system evidence boundaries, evidence-discipline guidance, and developer handoff work.","status_labels":[{"label":"current","meaning":"Published on the public site or machine-facing route today."},{"label":"next","meaning":"Launch-hardening work that should happen before broad public push."},{"label":"planned","meaning":"Adopted future work that needs more fixtures, tooling, or governance before it becomes a public claim."},{"label":"research-track","meaning":"Useful input that is not a current launch commitment."}],"today_focus":[{"id":"authority-orientation","status":"current","question":"What should a new reader understand first?","action":"Start with UAI-1 and Standards Fit, then use the roadmap before repeating broad claims about runtime fit, bridge profiles, SDKs, CLIs, adapters, certification, endorsement, or sync.","verify":["\/en-us\/specification\/uai-1\/","\/en-us\/specification\/standards-fit\/","\/en-us\/roadmap\/","\/wp-json\/uaix\/v1\/roadmap"],"boundary":"UAIX is the UAI-1 authority and portable evidence layer, not an agent runtime, scheduler, SDK, CLI, adapter program, certification body, or replacement for adjacent protocols."},{"id":"first-proof-run","status":"current","question":"How should an implementer produce one useful proof today?","action":"Choose one public profile, validate one packet, keep the result beside the Adoption Kit or Conformance Pack, and cite the changelog before treating the work as public support.","verify":["\/en-us\/tools\/validator\/","\/en-us\/tools\/adoption-kit\/","\/en-us\/tools\/conformance-pack\/","\/en-us\/governance\/changelog\/"],"boundary":"One passing validator result is packet evidence, not certification, endorsement, compliance, or ecosystem-wide support."},{"id":"ai-memory-package-work","status":"current","question":"What AI Memory work is useful and supportable today?","action":"Use the AI Memory Package Wizard to produce generated local files, package model JSON, manifest overlay JSON, visitor-AI digest output, copy-paste decks, .uai\/startup-packet.uai, .uai\/memory-maintenance.uai, .uai\/identity.uai, .uai\/world-context.uai, .uai\/totem.uai, .uai\/taboo.uai, .uai\/talisman.uai, and .uai\/progress.uai universal required files, active-anchor guidance for reading and obeying Totem, Taboo, and Talisman before broad execution, .uai\/coding-standards.uai and .uai\/next-recursive-prompt.uai required for Project Handoff packages, project\/developer packages, and packages that declare code, automation, executable workflow, or code-like artifacts, Multi-Agent Communication files required for Multi-Agent Communication configuration, Agent File Handoff plan and .uai\/intake-outcome-ledger.uai proof-of-use records required for Agent File Handoff configuration, the intake completion gate required for Agent File Handoff configuration, .uai\/long-term-memory.uai semantic pointer ledger output required for durable-memory configuration, and canonical starter ZIP links.","verify":["\/en-us\/ai-memory\/","\/en-us\/tools\/ai-memory-package-wizard\/","\/en-us\/guides\/agent-communication-operating-model\/","\/en-us\/specification\/project-handoff\/","\/wp-json\/uaix\/v1\/roadmap"],"boundary":".uai\/totem.uai, .uai\/taboo.uai, and .uai\/talisman.uai are active instruction anchors that must be read and obeyed before broad execution. The default .uai\/totem.uai carries the recursive Suggested Next Prompt contract: accepted work must end with a visible, copy-paste-ready Suggested Next Prompt containing literal Feature\/improvement focus, Code\/source changes, and Automated tests\/checks sections, concrete next work or explicit no-op\/source-validation stop condition, and enough detail for human review. Suggestions and permission requests are allowed; changing, deleting, weakening, replacing, or bypassing an anchor is forbidden unless a human names the exact artifact and exact operation. Managed package records, hosted upload\/import validation, hosted messaging, automatic synchronization, runtime orchestration, automatic repository writes, automatic LLM Wiki sync, SDKs, CLIs, certification, and endorsement remain unsupported until public evidence exists."},{"id":"agent-communication-operating-model","status":"current","question":"How should multi-agent communication be recorded without turning UAIX into a runtime?","action":"Use the Agent Communication Operating Model guide, eight uai.agent.*.v1 schemas, fixtures, validator checks, conformance cases, and AI Memory Wizard preset to record identity, intent, context, acknowledgement, evidence, memory proposals, deliberate promotion, and final handoff.","verify":["\/en-us\/guides\/agent-communication-operating-model\/","\/en-us\/specification\/schemas\/","\/en-us\/specification\/registry\/","\/en-us\/specification\/examples\/","\/en-us\/tools\/validator\/","\/en-us\/tools\/conformance-pack\/","\/en-us\/tools\/ai-memory-package-wizard\/"],"boundary":"Agent runtimes execute. UAIX records the reviewed communication, memory, trust, evidence, and handoff boundary. UAIX does not provide automatic sync, hosted messaging, runtime orchestration, official adapters, repository write execution, certification authority, or hosted import mechanisms."},{"id":"agentic-system-evidence","status":"planned","question":"How should agentic-system integration move forward?","action":"Keep runtime execution, tools, approvals, tracing, interruptions, orchestration, and managed memory with the chosen runtime while UAIX records portable UAI-1 evidence, Project Handoff memory, and release records that survive the run.","verify":["\/en-us\/specification\/standards-fit\/","\/en-us\/specification\/project-handoff\/","\/en-us\/guides\/project-handoff-openai\/","\/en-us\/roadmap\/"],"boundary":"Trace exporters, local redaction lint, reference adapters, and runtime-specific exporters stay planned or research-track until fixtures, owners, and release evidence exist."}],"documentation_ledger":[{"id":"document-roadmap-work","status":"current","title":"Document roadmap work where readers can verify it","summary":"Roadmap-changing work should update the public Roadmap, roadmap REST payload, canonical roadmap doc, focused tests, and compact handoff state together.","records":["\/en-us\/roadmap\/","\/wp-json\/uaix\/v1\/roadmap","docs\/roadmap.md","scripts\/test-uaix-ai-memory-bundle.php","AGENTS.md","readme.human",".uai\/progress.uai"],"evidence_rule":"A roadmap statement becomes reusable current-support language only when public page copy, matching machine artifact, reproducible evidence, implementation or package proof when needed, and dated release trail agree.","boundary":"Private chat, screenshots, archive folders, and unpromoted reports are not public roadmap truth by themselves."},{"id":"keep-source-leads-bounded","status":"current","title":"Keep intake files and reports as source leads until promoted","summary":"Active intake, Reports, and AIWikis cold memory can inform roadmap work, but their conclusions are source input until promoted into current UAIX records.","records":["agent-file-handoff\/Content\/","agent-file-handoff\/Improvement\/","\/en-us\/reports\/","AIWikis raw\/system-archives\/2026-05-01-uaix-internal-memory-reorg.json"],"evidence_rule":"Useful source facts become UAIX truth only when restated in current docs, public pages, machine artifacts, tests, release notes, roadmap state, or handoff records.","boundary":"Indexing, archiving, or preserving source material is not approval, certification, endorsement, validator evidence, conformance evidence, or permission to widen support claims."},{"id":"name-missing-proof","status":"current","title":"Name missing proof before support language moves","summary":"If a claim lacks a public page, machine artifact, reproducible fixture or check, implementation\/package proof when needed, or release trail, keep it next, planned, or research-track.","records":["evidence_promotion_gates","support_claim_checks","\/en-us\/governance\/launch-readiness\/","\/en-us\/governance\/changelog\/"],"evidence_rule":"Missing proof should be named in roadmap language instead of hidden behind broad support wording.","boundary":"Adapters, trace exporters, hosted imports, automatic repository writes, automatic sync, SDKs, CLIs, certification, and endorsement remain unsupported until the missing proof is public."},{"id":"future-agent-pickup","status":"current","title":"Make future agent pickup deterministic","summary":"Future agents should start from AGENTS.md, readme.human, loaded .uai files, direct active-bucket intake enumeration, docs\/roadmap.md, and the roadmap REST payload before changing support language.","records":["AGENTS.md","readme.human",".uai\/context.uai",".uai\/public-surface.uai",".uai\/decisions.uai",".uai\/progress.uai","\/wp-json\/uaix\/v1\/roadmap"],"evidence_rule":"Project truth changes should update the roadmap, machine payload, tests, changelog or release notes, and handoff state together.","boundary":"Future agents should not create competing instruction files, recurring live-site dependencies, CI pickup, watchers, scheduled jobs, or background handoff automation."}],"current_work_queue":[{"id":"hierarchical-goal-memory-guide","status":"current","title":"Publish Hierarchical Goal Memory guide and local-first wizard exports","summary":"Use source-linked subgoals, Folded Records, Derived Memory Hierarchies, bounded evidence retrieval, Context Packs, Fresh-Context Review, and review-gated promotion for long-running UAIX work without making generated summaries authoritative.","evidence_needed":["public Hierarchical Goal Memory guide route","AI Memory Package Wizard optional default-off export files","discovery and llms links to the guide","Playwright coverage for default-off wizard behavior and selected export files","changelog and .uai dogfood memory showing no hosted runtime, SDK, CLI, certification, automatic sync, automatic import, or repository-write claim"]},{"id":"agent-fallbacks-full-spectrum-support","status":"current","title":"Keep Agent Fallbacks and Full-Spectrum Support current","summary":"Require action-capable route records to name highest_supported_path, lowest_safe_fallback, POST\/live GET\/MCP\/auth\/tool fallbacks, human_review_url, and no_op_behavior so clients can ramp up only after capability is proven and ramp down before guessing.","evidence_needed":["public \/en-us\/spec\/agent-fallbacks\/ page","spec\/agent-executability-matrix.json with expanded route_record_required_fields","agent-compatibility.json and well-known discovery mirrors with the same fallback fields","validator-rule documentation for highest\/lowest path, MCP\/auth\/tool fallback, human review, and no-op warnings","llms.txt and llms-full.txt low-capability instructions that separate UAI-1 v1.0 from implementation-package versions"]},{"id":"dotnet-adaptive-interoperability-track","status":"planned","title":"Plan adaptive .NET package support for agent-to-agent interoperability","summary":"Extend the UAIX-owned .NET package line only after package code and tests prove capability-profile reading, fallback-chain resolution, ASP.NET Core route metadata helpers, typed HttpClient behavior, UAI-1 evidence export, and configuration-specific MCP\/A2A adapters where a host runtime provides those protocols.","evidence_needed":["NuGet package versions published and listed for each package ID before public install claims","unit and integration tests for ramp-up\/ramp-down path selection","ASP.NET Core middleware or route metadata examples that publish executable matrix fields","fixtures proving MCP\/A2A\/tool\/auth unavailable fallbacks without claiming hosted UAIX runtime support","docs that state .NET package versions do not change the UAI-1 v1.0 standards version"]},{"id":"deploy-current-artifacts","status":"next","title":"Deploy current packages and root assets","summary":"Upload or apply the freshly generated 2026-05-02 publish artifacts, then verify production static-root headers, HSTS, discovery files, and well-known manifests against the launch host.","evidence_needed":["deployed uaix-core.zip containing the well-known public-page URL normalization patch","deployed prod-fallback-root-assets with robots, sitemap, favicon, well-known files, and production .htaccess or edge equivalent","production response-surface check passing for public pages, REST routes, robots, sitemap, well-known manifests, HSTS, and static-root security headers"]},{"id":"serializer-parity","status":"next","title":"Prove AI Memory Package Wizard serializer parity","summary":"Show that one set of wizard selections produces equivalent generated local files, copy-paste decks, overlay JSON, visitor-AI digest summaries, and canonical ZIP references.","evidence_needed":["focused wizard regression fixtures","stale-review and changed-selection warnings before export","matching generated system profile, receiver brief, startup packet, long-term pointer ledger required for durable-memory configuration, overlay JSON, digest, and ZIP reference evidence"]},{"id":"local-validation-redaction","status":"planned","title":"Plan local validation and redaction lint","summary":"Define local package validation and redaction checks that help teams review AI Memory and Project Handoff files without implying hosted upload\/import validation or automatic repository writes.","evidence_needed":["public fixtures for allowed and blocked handoff content","documented redaction and source-authority rules","release notes that keep local lint separate from certification, endorsement, SDK, or CLI claims"]},{"id":"trace-to-handoff-fixtures","status":"planned","title":"Design trace-to-handoff evidence fixtures","summary":"Map selected runtime traces, tool calls, approvals, task status, and handoff summaries into UAI-1 records or Project Handoff files only after public examples, review gates, and redaction rules exist.","evidence_needed":["sample runtime-to-UAI evidence packets","redacted trace identifier and approval posture examples","validator expectations or handoff checks that preserve the runtime boundary"]},{"id":"capability-package-sidecars","status":"research-track","title":"Evaluate capability-package sidecars and mapping annexes","summary":"UAIX publishes a reference UAIX \/ UAI-1 Agent Skill bundle as reusable procedure guidance, while broader selected UAI evidence mappings for skill folders, plugin manifests, action groups, agent cards, MCP\/OpenAPI capability surfaces, and subagent profiles remain research-track without treating discovery as authorization or current package-profile support.","evidence_needed":["keep the reference Agent Skill bundle scoped to procedure loading for creating, validating, citing, packaging, and handing off UAI-1 evidence","source-verified mapping examples for SKILL.md packages, A2A Agent Cards, MCP\/OpenAPI tool surfaces, plugin manifests, declarative-agent manifests, and action groups","package identity, semantic version, native reference, approval mode, allowed tools, sandbox profile, provenance, integrity digest, owner, distribution scope, review state, and stale-review fields drafted as selected sidecar metadata","public fixtures, validator or handoff checks, redaction and trust-boundary guidance, ownership, maintenance scope, and release notes before any package profile becomes current support"]},{"id":"reference-adapter-evaluation","status":"research-track","title":"Evaluate reference adapters after evidence exists","summary":"Study OpenAI, A2A, MCP, and runtime-harness adapter patterns as evidence exporters, not as UAIX-owned execution, orchestration, or official vendor-adapter support.","evidence_needed":["adapter scope statements that leave execution with the runtime","fixture-backed export examples","maintenance ownership before any adapter leaves research-track status"]},{"id":"trust-metadata-expansion","status":"current","title":"Expand trust metadata after error-profile harmonization","summary":"Source-ready UAI-1 trust metadata now keeps the six-field compact\/keyless trust prefix and adds verifier posture, credential posture, proof references, replay-window policy references, verification timestamps, and assurance-level descriptors not required for compact trust readers without forcing one credential stack.","evidence_needed":["normal package build and deployment path before production routes or ZIP artifacts are described as updated","continued support-boundary checks that avoid hosted credential validation, certification, official adapter, or runtime-enforcement claims"]}],"evidence_promotion_gates":[{"gate":"public-page-update","required_before_current":"The current-support wording appears on the relevant UAIX public page with support boundaries and route links."},{"gate":"machine-artifact-update","required_before_current":"The matching JSON, schema, registry, roadmap, discovery, package, or conformance artifact exposes the same state."},{"gate":"fixture-or-validator-evidence","required_before_current":"Examples, fixtures, validator expectations, package checks, or handoff lint make the behavior reproducible."},{"gate":"implementation-or-package-evidence","required_before_current":"A named implementation track, starter package, conformance packet, or release artifact carries the proof."},{"gate":"dated-release-trail","required_before_current":"The changelog, news, roadmap state, or durable handoff record explains what changed and what remains unsupported."}],"support_claim_checks":[{"id":"current-support","question":"Can this be described as current UAIX support?","answer":"Yes only when the relevant public page, matching machine artifact, reproducible fixture or validator\/package evidence, implementation or package proof when needed, and dated release trail all agree.","default_if_missing":"planned-or-research-track","verify":["Roadmap status is current","canonical public page carries the same boundary","matching machine artifact or package exposes the same state","validator, fixture, package check, or handoff lint makes the behavior reproducible","changelog, news, roadmap state, or durable handoff record records the change"]},{"id":"ai-memory-package-scope","question":"Can UAIX claim managed AI Memory packages, hosted upload\/import validation, automatic repository writes, or automatic LLM Wiki sync?","answer":"No. Current support is the guided AI Memory Package Wizard, canonical starter ZIP links, generated local files, copy-paste decks, package-model JSON, manifest overlay JSON, visitor-AI digest, generated system profiles, receiver briefs, startup packets, Multi-Agent Communication files required for Multi-Agent Communication configuration, and long-term semantic pointer ledger output required for durable-memory configuration.","default_if_missing":"unsupported-until-evidence","verify":["AI Memory Package Wizard public page","canonical bundle registry and starter ZIP routes","wizard serializer parity work queue item","non-claim boundary for hosted import, automatic writes, automatic sync, SDK, CLI, certification, and endorsement"]},{"id":"agent-communication-scope","question":"Can UAIX claim hosted messaging, automatic synchronization, runtime orchestration, official adapters, repository write execution, certification authority, or hosted import mechanisms?","answer":"No. UAIX publishes the Agent Communication Operating Model, schemas, fixtures, validator behavior, conformance cases, profile pages, examples, and selected portable wizard files. Agent runtimes execute. UAIX records the reviewed communication, memory, trust, evidence, and handoff boundary.","default_if_missing":"unsupported-until-evidence","verify":["Agent Communication Operating Model guide","eight uai.agent.*.v1 schemas and fixtures","validator policy checks for identity, UTC, expiry, secrets, cold memory, blockers, and final reports","conformance pack regression cases","AI Memory Package Wizard Multi-Agent Communication preset","Changelog and launch note support boundary"]},{"id":"adapter-and-runtime-scope","question":"Can UAIX claim official adapters, runtime execution, orchestration, tool access, approvals, tracing, or managed memory?","answer":"No. UAIX can describe portable UAI-1 evidence and Project Handoff memory around runtime systems, while the chosen runtime keeps execution, tools, approvals, tracing, interruptions, orchestration, and managed memory.","default_if_missing":"research-track","verify":["Standards Fit agentic-system boundary","Project Handoff runtime boundary","trace-to-handoff fixture planning","reference-adapter evaluation queue item","release evidence before any official adapter, SDK, or CLI claim"]},{"id":"conformance-claim-scope","question":"Can one passing validator result be repeated as certification, endorsement, compliance, or ecosystem-wide support?","answer":"No. A passing validator result is evidence for one reviewed packet, not certification. Broader support language needs the validator result, conformance pack, implementation evidence, fixture coverage, public page copy, and release trail to agree.","default_if_missing":"packet-only-evidence","verify":["Validator output","Conformance Pack","implementation evidence checklist","fixture coverage","Roadmap and Changelog support boundary"]}],"priorities":[{"id":"production-hardening","status":"next","title":"Production launch hardening","summary":"Keep deployment packaging, root discovery files, sitemap delivery, security headers, and launch audits aligned before wider publication.","evidence":["Launch Readiness page","scripted WordPress ZIP publishing","root robots.txt, sitemap.xml, and sitemap.html","content-type audit checks for root JSON and XML discovery assets","query-string alias redirect checks for clean locale-prefixed canonical URLs","locale-prefixed public routes","security-header baseline on WordPress-rendered responses","launch-surface audit checks for public machine routes and response headers","zh-CN rendered-route translation audit in the publish path"]},{"id":"standards-fit","status":"current","title":"Standards-fit explanation","summary":"Explain the relationship with A2A, MCP, OpenAPI, JSON Schema, DID\/VC, Trace Context, JCS, compact transfer, Problem Details, and agentic runtime harnesses without claiming replacement of those systems.","evidence":["Standards Fit page","UAI-1 protocol-fit page copy","quick chooser for UAI-1 vs MCP and UAI-1 vs A2A decisions","agentic systems architecture path that separates runtime execution, tool access, approvals, tracing, and durable UAIX evidence","adjacent-standards references","validator-backed bridge evidence examples for A2A, MCP, OpenAPI, DID\/VC, and Trace Context","compact-transfer ladder and normalization rule"]},{"id":"public-operating-layer","status":"current","title":"Public operating layer","summary":"Publish a dedicated contact-and-review path for contributor packets, change proposals, decision-trail notes, release cadence, and review intake guidance while broader issue trackers remain future work.","evidence":["Contact and Review page","References and Contributors handoff","Governance page","Changelog release discipline"]},{"id":"project-context-handoff","status":"current","title":"Project context handoff","summary":"Publish draft AGENTS.md plus .uai portable project-memory guidance, AI Memory starter bundles, the AI Memory Package Wizard, OpenAI\/Codex compatibility guidance, generated long-term pointer-ledger guidance, AGENTS.md linking and LLM Wiki comparison report material, active local file-intake guidance, a permanent report trail, required first-response guidance for new AIs, public route guidance into UAI-1 evidence, and loader\/security guardrails. Keep UAIX as the evidence and handoff layer around agentic systems while hosted upload\/import validation, automatic repository writes, automatic LLM Wiki sync, OpenAI adapters, trace-to-handoff exporters, SDKs, CLI tools, certification, and endorsement remain planned.","evidence":["AI Memory page","AI Memory Package Wizard page","canonical AI Memory bundle registry","package model JSON, manifest overlay JSON, generated system profiles with source-authority, evidence-ledger, conflict-resolution, risk, and rollback protocols, generated receiver briefs, generated startup packets, copy-paste file decks, exact-root long-term semantic pointer ledger files required for durable-memory configuration, readiness metadata, and canonical starter ZIP links","Project Handoff page","Using UAIX Project Handoff with OpenAI Agents and Codex guide","Agent File Handoff page","AGENTS.md .uai Linking Specification page","Guides index with an agentic system evidence path","Reports index with evidence-discipline guidance","LLM Wiki vs. UAIX Project Handoff report","LLM Wiki and UAIX Project Handoff report","Refining UAI Protocol for Agent Communication report","required first-response guidance before broad AI edits","AIWikis.org dogfood correction requiring generated packages to include active agent-file-handoff folders and a fuller .uai handoff set","public route guidance from Project Handoff into UAI-1, schemas, registry, examples, validator, Adoption Kit, Conformance Pack, Roadmap, and Changelog","launch-surface audit coverage for project-handoff, agents-md, reports, LLM Wiki report, and refining-report pages","public support boundary that separates draft repository-context files from UAI-1 conformance evidence","local intake boundary that keeps dropped files out of public production content until review","agentic-system boundary that leaves execution, tool access, approvals, tracing, interruptions, and orchestration with the chosen runtime"]},{"id":"ai-memory-package-authoring","status":"planned","title":"Governed AI Memory package authoring","summary":"Move beyond the current guided wizard only after one canonical package object model, serializer parity, review states, provenance records, privacy gates, stale-review checks, artifact versioning, and reviewed AIWikis archive sync have implementation evidence and operations ownership.","milestones":[{"status":"current","title":"Guided package planning","criteria":["eight-step browser wizard over the supported bundle registry with the intake completion gate required for Agent File Handoff configuration","package model JSON, manifest overlay JSON, generated system profiles, receiver briefs, startup packets, long-term semantic pointer ledger files required for durable-memory configuration, copy-paste file decks, readiness metadata, and canonical starter ZIP links","visitor-AI digest generated from the same canonical wizard payload"]},{"status":"next","title":"Serializer parity and review-preview hardening","criteria":["prove the same package selections produce equivalent copy-paste decks, generated local files, overlay JSON, AI digest summaries, and canonical ZIP references","make stale-review and changed-selection warnings explicit before export","keep the result as local planning output, not hosted upload\/import validation"]},{"status":"planned","title":"Managed package records","criteria":["define package-instance identifiers, review states, provenance records, privacy and redaction gates, artifact versioning, and ownership rules","publish fixtures and operations guidance before managed package records become current support","keep automatic repository writes, WordPress admin installation, SDKs, CLIs, certification, and endorsement out of current support"]},{"status":"research-track","title":"Adapter and archive-sync experiments","criteria":["evaluate OpenAI adapter and trace-to-handoff exporter patterns only after package records have evidence","evaluate reviewed AIWikis archive-memory sync as an explicit reviewed export, not automatic bidirectional sync","keep research-track experiments out of public support language until public fixtures, ownership, and release evidence exist"]}],"evidence":["current AI Memory Package Wizard as guided package-planning UI","current canonical bundle registry and deterministic starter ZIP routes","current visitor-AI digest generated from the same canonical wizard payload","next serializer parity across generated files, copy-paste deck, overlay JSON, AI digest, and ZIP references","next stale-review and changed-selection warning behavior","planned package-instance model","planned serializer parity across copy-paste, ZIP, and managed records","planned review-state and provenance metadata","planned privacy and redaction gates","planned reviewed-export sync into AIWikis archive memory","planned OpenAI adapter and trace-to-handoff exporter as future compatibility work","non-claim boundary for hosted upload\/import validation, automatic repository writes, SDK, CLI, certification, and endorsement"]},{"id":"agent-file-handoff","status":"current","title":"Agent file handoff and active intake","summary":"Dogfood an AGENTS.md-triggered local active-bucket intake pattern so files supplied by humans, other AI systems, or adjacent tools become visible during normal project handoff load, regardless of prompt wording, without becoming public truth automatically; already-dispositioned files move to an archive that routine AI intake ignores.","evidence":["Agent File Handoff page","root AGENTS.md loaded-context guidance for .uai\/file-handoff.uai and direct active-bucket enumeration","agent-file-handoff\/Content and agent-file-handoff\/Improvement active intake buckets","agent-file-handoff\/Archive dead archive for already-dispositioned files","Content and Improvement directory listings as the pending-intake source of truth","Project Handoff WordPress prototype export scaffolding for portable intake\/archive folders, fuller .uai package context, and no bucket-local README guidance","LlmWikis.org dogfood workspace consuming the same bucketed intake and dead archive shape"]},{"id":"conformance-maturity","status":"next","title":"Conformance maturity","summary":"Keep the validator normalization modes, regression fixtures, canonical-hash evidence, route\/cache posture, trust-boundary evidence, reader QA, result exports, and CI-ready conformance packets expanding before making broader assurance claims.","evidence":["uai.conformance.result.v1 records","conformance pack","expanded public implementation-evidence checklist covering identity, profile scope, validator evidence, machine-route\/cache posture, trust and threat boundary, locale\/accessibility reader QA, fixtures, release trail, and support boundary","reusable positive and negative conformance fixture pack with traceparent, DID\/VC, required-field, undeclared-field, keyless-overflow, and unsupported-alias boundaries","bridge evidence pack with validated A2A, MCP, OpenAPI, DID\/VC, and Trace Context mapping examples","validator workbench","current keyed, minified-keyed, and keyless validator normalization","canonical hash equivalence across keyed, minified-keyed, and keyless inputs","published keyed and keyless example payloads","future alias and binary fixture coverage"]},{"id":"compact-transfer","status":"current","title":"Compact-transfer discipline","summary":"Keep keyless JSON tied to the field registry and normalize it back to keyed JSON before validation, hashing, signing, or public conformance evidence; future alias or binary transport formats still need fixtures.","evidence":["field-registry route","keyed and keyless examples","validator-backed keyless expansion","JCS-compatible hash over reconstituted keyed JSON"]},{"id":"developer-handoff","status":"planned","title":"Developer handoff","summary":"Move from page-led onboarding, supported starter downloads, and guided wizard outputs toward local validation, canonical package serialization, reference adapters, trace-to-handoff exporters, reusable SDK, CLI, source-repository, hosted generator, import validator, managed package, and local harness material only after the public artifact set and maintenance model are explicit.","evidence":["AI Memory","AI Memory Package Wizard","Project Handoff","AGENTS.md .uai Linking Specification","API Reference","Adoption Kit","implementation evidence checklist","conformance fixture pack","bridge evidence pack","OpenAPI route","Conformance Pack","planned Python and TypeScript helper evaluation","planned A2A and MCP wrapper examples","planned reference adapters that map runtime harness evidence into UAI-1 records without taking over runtime execution","planned trace-to-handoff exporters and redaction lint once fixtures, ownership, and release evidence exist","planned NIST\/IETF crosswalk and regulated-evidence template evaluation"]}],"interoperability_adjacency":[{"name":"A2A","fit":"Agent discovery, delegation, and task coordination can sit beside UAI-1; UAI-1 carries the portable public exchange and evidence record."},{"name":"MCP","fit":"Host-client-server tool sessions and application-local capability negotiation can stay in MCP; UAI-1 carries reviewable exchange records that need to travel outside that boundary."},{"name":"OpenAPI","fit":"OpenAPI describes HTTP APIs; UAIX publishes an OpenAPI document for the REST surface while UAI-1 describes the exchange message contract."},{"name":"DID\/VC","fit":"DID and Verifiable Credential ecosystems can support trust.principal, credential_ref, and signature_ref without becoming a single mandatory UAIX identity stack."},{"name":"W3C Trace Context","fit":"traceparent can travel in conversation metadata when distributed tracing already exists."},{"name":"RFC 9457 Problem Details","fit":"Problem Details informs the typed public error style for uai.error.v1 while UAIX keeps the profile-specific error registry."}],"normalization_modes":[{"mode":"keyed-json","status":"current","rule":"Readable keyed JSON is the reviewable source record."},{"mode":"minified-keyed-json","status":"current","rule":"Whitespace removal does not change the normalized keyed record or canonical hash."},{"mode":"keyless-json","status":"current","rule":"Keyless records reconstruct through the public field registry before validation, result export, and canonical hashing."},{"mode":"alias-json","status":"planned","rule":"Alias-key transport remains future work until aliases, fixtures, normalization, and validator expectations are published."},{"mode":"binary-envelope","status":"research-track","rule":"Binary transfer remains background exploration until encode\/decode parity fixtures and public route behavior exist."}],"metrics":["public conformance packet count","positive and negative conformance fixture coverage","normalization mode coverage","canonical hash fixture coverage","traceparent and DID\/VC trust-boundary fixture coverage","required-field and undeclared-field regression coverage","bridge evidence example coverage and future formal bridge-profile fixture coverage","implementation evidence checklist completion","project-context handoff route coverage, AI Memory Package Wizard serializer parity, generated system profile coverage, generated long-term pointer-ledger\/readiness coverage, agent-file-intake coverage, and loader-guardrail coverage","agentic-system evidence coverage for runtime-to-UAI record exports, trace-to-handoff experiments, redaction lint, and canonical package serialization once those artifacts exist","report-to-current-record promotion coverage for strategy inputs that become public claims","managed package authoring review-state and provenance coverage once that work exists","implementation-track evidence count","byte-size delta between keyed, minified, keyless, alias, and binary candidate formats","launch-readiness gate completion across response, package, accessibility, locale, and release-trail checks","response-header parity across public HTML, REST, and static root discovery assets","content-type parity for root JSON, XML, robots, and sitemap discovery assets","release-note completeness for every public artifact change"],"non_claims":["UAIX is not a certification body today.","UAI-1 does not replace A2A, MCP, OpenAPI, identity, signing, tracing, or transport systems.","UAIX does not replace agent runtimes, model execution, tool access, approvals, tracing, interruptions, orchestration, or managed memory platforms.","Current bridge evidence examples are mapping examples, not completed bridge profiles, SDK support, or certification claims.","Project Handoff, Agent File Handoff, and AGENTS.md-linked .uai files are draft repository-context guidance, not UAI-1 conformance evidence, certification, or endorsement by themselves.","The AI Memory Package Wizard is a guided package-planning UI over supported starter bundles with generated system profiles that now include source-authority, evidence-ledger, conflict-resolution, risk, and rollback protocols, generated receiver briefs, generated startup packets, and the intake completion gate required for Agent File Handoff configuration; it is not hosted upload\/import validation, automatic repository writing, automatic LLM Wiki sync, SDK, CLI, certification, endorsement, or WordPress admin package management.","Capability-package sidecars and mapping annexes are research-track evidence-mapping ideas, not a current package profile, registry, authorization layer, official adapter, SDK, CLI, certification, endorsement, or conformance claim.","Alias and binary transport formats are not current public support until validator-backed fixtures and route behavior are published.","One passing validator result is evidence for that reviewed packet, not a blanket ecosystem support claim."],"route_url":"https:\/\/uaix.org\/wp-json\/uaix\/v1\/roadmap","download_url":"https:\/\/uaix.org\/wp-json\/uaix\/v1\/roadmap?download=1","public_page_url":"https:\/\/uaix.org\/en-us\/roadmap\/"}