Files
knowledge-wiki/brain.md
T

31 KiB
Raw Blame History

🧠 Knowledge Wiki — Brain Summary

Last updated: 2026-10-24 daily sync

  • Skills Public: RootAtSkic requested source analysis of Anime.js skills to improve presentation-html, grouped example registries under references linked from SKILL.md, and directly openable HTML demos under assets for each group. Latest human anchor 1563199351461904405; implementation, publication and user acceptance not independently revalidated by daily sync. Monitoring now contains 38 IDs, including Skills Public; 33 installed skill/wiki routes verified (maximum 262 characters).

  • CUGA presentation refinement: RootAtSkic requested HTML-presentation skill routing, ambitious animations beyond panel pop-ups, continued refinement, self-testing and animation across all slides. Balanced comparison remains governing; implementation not independently validated by daily sync. Anchor 1562882559682158643.

  • Conduit onboarding: keep dashboard authentication enabled, use an iPhone-reachable URL and provide a QR. Channel workflow records password login alongside SSO and private DM delivery; device acceptance remains unverified. User also challenged new-skill Gitea naming/ownership compliance; resolution unknown. Never mirror credentials or QR payloads. Anchor 1562747828982521897.

  • Wiki coverage gap (2026-10-23): 32 mapped routes validate, but routed CUGA delivery 1562800574892744835 is not in the authoritative 37-ID scan mapping and was not scanned. Retained test 1518733120831226028 remains absent from live text inventory. No monitoring membership or response scope changed.

  • Monthly CogArch workshop (2026-10-22 sync): current task covers Shorts introduction, Upcoming Shorts and Reference Architectures with five in-progress pilots/search reminder. User reported actual PPTX rendering unlike previews; existing correction/LibreOffice record is not final PowerPoint/Keynote acceptance. Anchor 1562569220850581606.

  • CUGA comparison (2026-10-22 sync): balanced standalone Hermes–CUGA evaluation based on public sources, not a migration/adoption argument; retain three final feature-mapping appendices. Initial history reconciled through 1562435905346609214; no deployment/integration authorized. Channel-workflow delivery evidence retained without independent rerun.

  • Eldorado repository (verified 2026-10-22): user requested public visibility; authenticated metadata confirms eldorado-v1/documentation is public. Confirmation anchor 1562405021952245894; no settings changed by daily sync.

  • eShop Shorts direction (2026-10-17): RootAtSkic requested topics for a connected series building eShop architecture before implementation (1561039420575064177). Episode sequence, technology stack and distribution remain undecided; no production/publication approval inferred. Latest processed human anchor 1561045336582004757; live daily read verified 2026-10-18.

  • Kreg/Ingrida presentation (2026-10-14): Ingrida JL (1465343341435686985) requested a deck based on supplied DOCX/PPTX with little slide text; RootAtSkic required reading the history, responding to her, and following up on the new presentation. Existing 17-slide local delivery record is channel-workflow evidence, not fresh acceptance/native-render verification. Latest sandbox anchor 1559894132892110941; general responsiveness anchor 1559828160634163242.

Workstreams & Status

# Workstream Open Items Priority
🔵 VSSA — DPVP 16 active, 1 opportunity Phase 2 opportunity, Azure-cost onboarding documentation, access governance/monitoring, Jira-targeted role grants/revocations, unit-testing options, and cloud-role review
🟡 VSSA — Enterprise Architect 4 active Scope decision, Gytis outreach, Atea & IBM institution engagement, Government Developer Portal
🟢 Cognitive Architect 8 active Oleg follow-up; Aurimas Nogis/Hermes security follow-up; Texture/Jenny integration follow-up; Anthropic architect certification; IBM Video Streaming; IBM Method review; cogarch backlog
🟣 SAP Delivery Powered By AI 1 active Michael Martine material review
🏅 Distinguished Engineer 3 active Elizabeth docs + summary; FutureNow/Baltic AI/client-relations upskilling with Jurgis

Total open: 33 items | Completed: 7 (env vars deployed; Apptio AWS-connection investigation and Azure-cost onboarding completed; CA-01, CA-03, CA-04, and CA-B01 completed)

Atea Cooperation

  • #atea-storm is the continuity ledger for Atea-related contracts, opportunities, audits, and RFPs.
  • Current tracked workstreams: VSSA AWS contract; VSSA DPVP contract and Phase 2 opportunity; Valstybės duomenų agentūra AWS portal audit; Lietuvos Geležinkeliai RFP; and two early Atea & IBM items (the user-provided title “Let's go to institutions” and a Government Developer Portal). Details for the two Atea & IBM items are not yet specified.
  • DPVP Phase 2 discussion scope covers the portal, backend execution migration, VSSA on-premises onboarding, complex hybrid services, and organization scanning across AWS, Azure, and GCP. Ownership, prioritisation, estimates, dependencies, dates, scan baselines, and commercial route remain unconfirmed.
  • VSSA AWS: an AWS Well-Architected Framework review has been proposed, not approved; workload/account scope, authorization, access, schedule, and remediation ownership remain open.
  • DPVP backlog changes confirmed 2026-08-25: the Apptio AWS-connection investigation and Azure-cost onboarding are completed; the follow-up with Atea about TFE OIDC role assumption and July cost-estimation review were dropped. Open refinements include figuring out AWS tagging issues, documenting the Azure-cost onboarding approach, and having Julius locate documentation for the nightly AWS cost-allocation-tag enablement job.
  • The portal audit is an opportunity only; active testing requires explicit written authorization and agreed scope.
  • A consolidated executive summary was requested on 2026-07-29 and is maintained in wiki/channels/1531661088209371198.md.

VSSA Client Investigation

  • Current delivery (2026-09-04): repository vssa-v1/documentation publishes the Keycloak-protected VSSA - Clients & Systems Docusaurus site. The latest completed delivery records 290 clients, 551 canonical systems, 779 observations, and 118 contractor cards at commit 701494d175266bd5e6683f53c58d08b7cd69bc0e; exact-head Actions task 1874 succeeded. RootAtSkic stopped and removed continuous-research job 88833e00b4a2 on 2026-09-04, so it no longer runs every two hours. The unassigned next range remains clients 121–140.
  • Entity classification (confirmed 2026-08-19): Valstybės skaitmeninių sprendimų agentūra (VSSA) is a central-government client body, not a system or commercial contractor. Government bodies that centrally operate/provide shared services must not be converted to CTR-* entities merely because of that role.
  • Content organization (confirmed 2026-08-12): after Lego requested an immediate research-led overall-page improvement cycle, commit 1542674 reorganized the portal around researched user tasks. Navigation is Overview, Clients, Systems, Methodology, and Data; deployment was validated by successful Actions run 2335 and representative HTTP-200 backend/browser checks.
  • Evidence/navigation requirements (confirmed 2026-08-12/14): Overview, Clients, Systems, Contractors, Methodology, and Data have dedicated sidebars; Clients, Systems, and Contractors each expose Overview plus an expanded Registry. Internally generated SYS-*/CTR-* identifiers are documented, all generated client/system dossiers have References indexes, relationship evidence uses reusable cards, every client system observation links to its canonical system dossier, and every named contractor on an exact client-system row links to its Contractor Card. Multiple contractors are supported; unknown contractors remain explicit and unlinked.
  • VSSA skill governance (confirmed 2026-08-14): vssa-clients is authoritative in vssa-v1-skills-code-agent/vssa-clients; merged documentation-docusaurus is authoritative in vssa-v1-skills-code-agent/documentation-docusaurus and includes the retired gitea-docusaurus-projects capabilities. Both runtime skills are symlinked to Gitea maintainer worktrees. The requested corp-v1-skills-code-agent/documentation-docusaurus repository was unavailable and was not touched.
  • Procurement research (confirmed directive and validated workflow, 2026-08-12): Discord user 476287310627864587 is the designated instructor for CVP IS. In addition to tender search, use exact organization search → identity-checked organization profile → PERŽIŪRĖTI VISUS PASKELBTUS SKELBIMUS, extracting the actual authorityId/orgGroupId and relevant notice PDFs. Job 88833e00b4a2 must map supported records to exact clients/systems and keep RFI/RFP/specification/bid stages distinct from award, implementation, production use, hosting, and completion.
  • Lifecycle evidence rule (confirmed 2026-08-12): a dated official announcement explicitly stating that a named system started operating/launched/went live is direct production-release evidence at the announcement’s stated precision; generic webpage metadata is not. Do not infer delivery kickoff from a production announcement.

Key People

  • Lego (root_at_skic) — Owner, Enterprise Architect
  • MartynasP (w4rl0ck_21) — Hermes migration lead
  • Jerry — TrueNAS admin
  • Julius Sulcas — DPVP, was on time off
  • Rimvydas — Architect, reports to Dimitrian via Edvinas
  • Gytis Račiukaitis — E-Gov Gateway contact
  • Luis Diego Raga — Pentest senior tester (Sep 1–28)
  • Oleg — Cognitive Architect stakeholder
  • Andrius — Built cogarch skill
  • Elizabeth — Distinguished Engineer track
  • Jurgis — FutureNow, Baltic AI, client-relations, and broader upskilling contact
  • Jenny — Mentioned Texture integration and other Cognitive Architect items; exact items and role still need clarification
  • Michael Martine — SAP material contact
  • Arunas Saltis — DNS naming convention (internal hosting)

Pending Reminders

  • Follow up with Arunas Saltis — DNS naming convention (internal hosting) — added 2026-07-14
  • Check the theorem about the uncountability of logics — added 2026-09-23

Infrastructure

  • Kreg channel renames (verified 2026-10-14): former Emplonet general/Ingrida channels are now #kreg-v1-general (1546854666854334545) and #kreg-v1-sandbox (1546855326136148060). RootAtSkic requested matching local folders; both live names and matching /opt/data/channels/ directories were verified. Existing Emplonet skill/repository identifiers remain unchanged; the new domain scope is unspecified. Monitoring still uses correct stable IDs under old display labels. Human anchors: 1559581079378333737 / 1559580995223552094.

  • Shorts approved refinements (2026-10-11): accept blog-review points 1/2 (AI assistance versus human authority; synchronized views/model-level impact versus team review), leave conversion point 3 unchanged. Architecture-lead Markdown prompted approved role → applicability → evidence revisions to slides 10–13: five universal adoption considerations plus sixteen role-relevant criteria, no numerical scores/exclusive categories, platform versus tooling/team responsibilities. Slides 9/14 stay unchanged; latest human anchor 1558896305114320937. Channel workflow records v2.8.14 delivery; daily sync did not independently validate the deck or native Keynote import.

  • Shorts consistency review (2026-10-10): RootAtSkic requested checking the latest presentation against the Cognitive Architect blog and reporting contradictions (1558592734196138078). Blog access, comparison findings and completion remain unverified by daily sync. Keynote open failure was reiterated (1558440086649962618); recorded v2.8.12 is a compatibility candidate, not confirmed native-import success.

  • Eldorado presentation corrections (2026-10-09): remove slide-2 assignment/unknown labels; explain AI Agent with familiar examples; alter slide-7 layout without changing process order; clarify the weekly-template framing. Measurement needs concrete numeric quantities grounded in the existing Measurement page, not generic “Accepted work.” Latest anchor 1558010932209262632; implementation not established by daily sync.

  • Shorts delivery/compatibility (2026-10-09): integrate slide 9 into the full, consistently styled deck; place a taller Cognitive Architect foundation below the columns, reduce other fonts and correct unintended black backgrounds. User reported Keynote import failure; channel workflow records v2.8.11 corrections, but native Keynote opening remains unverified. Latest anchor 1558200114269659137.

  • Eldorado governing direction (2026-10-08): preserve long-term Target and Roadmap as separate pages alongside 6 Months Target, 6 Months Roadmap and Beyond; the assignment-driven six-month plan must not replace the long-term vision. Distinguish workflow/organization effectiveness from skills-health metrics, connect the entire strategy story, and align the documentation-styled Reveal.js presentation with it. Homepage embed heading: Understand In Speed. Departments become sidebar-visible Structure children. Anchor 1557830033090027611; detailed channel-workflow publication evidence is preserved but not independently revalidated by daily sync.

  • AeroSim 3D-model channel (2026-10-08): dedicated workspace and Gitea-backed skill requested in 1557702471294521394 (anchor 1557702572112879657). Existing local wiki/runtime route reconciled into daily monitoring: 36 authoritative IDs and 30 mapped skill routes; remote setup details remain channel-workflow evidence.

  • Eldorado Measurement/Baseline clarification (2026-10-07): Measurement goes first as an optional registry of candidate KPIs/metrics, not a mandate to collect or baseline every metric; a separate Baseline phase follows. RootAtSkic authorized the discussed updates (1557373230434754573, clarification 1557370994006560821). Channel workflow records Measurement → Baseline → Governance → Quality → Adoption → Maintenance and publication; daily sync reconciled human direction without independently revalidating deployment. This supersedes earlier mandatory-collection interpretations.

  • Eldorado strategy (2026-10-06): two stages—standardized skills with humans + AI Agents across SDLC phases, Delivery first, then a future Autonomous Agent. Each phase repeats current-state discovery → KPIs → skill topology → usage → feedback → renewed discovery; remaining phase order is undecided. Approved Stage 01/02 navigation; wider topic-specific diagrams, every page sidebar-visible, and each metric on its own changed canonical URL under Measurement. Anchor 1557166067191644183; implementation/publication claims in the channel ledger were not revalidated by daily sync.

  • VSSA Criminal Police inquiry (2026-10-06): RootAtSkic requested confirmation of “Kriminaline policija” as a VSSA client and its systems. Exact legal entity/client relationship/system inventory remain unknown in the extracted human evidence. Anchor 1556982849335459841.

  • Eldorado strategy refinements (2026-10-05): Target → Roadmap navigation; human-controlled autonomous agent at the center of an eight-phase SDLC cycle with external human-controlled AI Agents. Restore inline theme/style/export controls and portal-theme synchronization. Roadmap begins with code-quality/process baseline; Skills Measurement requires concrete adoption KPIs and collection/calculation guidance. Documentation links only on deployed-change reports. Latest anchor 1556776126842667058; deployment and final roadmap visual remain unverified by this scan.

  • Shorts current direction (2026-10-05): Architecture in AI World now uses AI Amplification → evidence → architecture capabilities/definition → unaligned-change example → human/AI shared-architecture response, followed by framework selection, linked cards, grouped evaluation criteria and Cognitive Architect focus (variant A). Keep reports alongside industry voices, concise embedded presenter notes, previous-Short recording and final Links; restore familiar Shorts title-page styling. Anchor 1556767129901072385; this is the superseding human direction, not fresh deck-validation evidence.

  • Cognitive Architect skill source/access (2026-10-05): user requested replacing the installed skill from github.ibm.com/CTOTools-skills-code-agent/cognitive-architect and testing architecture visibility via that skill, using refreshed Bitwarden COGARCH_AUTH_TOKEN/COGARCH_BASE_URL. Completion/access remain unverified by daily sync. Anchor 1556637831106007131.

  • Eldorado AI-adoption direction (2026-10-04): diagram-first, low-reading-load strategy tied to the defined organization. First focus is skills governance, quality, adoption, measurable KPIs and maintenance; long-term TO-BE is one SDLC-governing AI agent with human laptop code agents retained. Use only the dedicated Eldorado - Storm board, not AeroSim. Department layout, Collaboration/SDLC navigation, Contribution Overview and icon-only diagram controls are recorded in the channel wiki. Anchor 1556269523701727353; target architecture is a request, not a deployed-state claim.

  • Shorts current direction (2026-10-04, supersedes October 1 topic pause): resume Architecture in an AI World; tell a coherent story for architecture's increased importance, starting with a neutral question, retaining intent/design/decisions/delivery and ending with CALM/Structurizr/ArchiMate support. Retain AI's amplification of organizational strengths/weaknesses without presenting it as direct proof of the thesis. Eight-slide working outline and overnight research are drafts; no deck edit/publication is established by the daily scan. Anchor 1556388047036096655.

  • Eldorado Powered By AI (2026-10-02): new monitored channel 1555510466904727622 owns a dedicated eldorado-storm-v1 skill and workspace. Its channel ledger records private eldorado-v1/documentation, a Docusaurus/Pages foundation with Organization, Strategy, Contribution navigation and Archify diagrams. Human direction: same-page fullscreen diagrams; department-owned, non-shared teams, with exactly two in Product and five in Technology (confirmed 2026-10-03, decision 1555831103577784421, follow-up 1555831136339628163). This resolves the previous allocation ambiguity. The organization chart is an operating model, not independently certified staffing; CEO/C-level research remains open in the extracted human evidence. Latest message 1555831136339628163; repository/deployment claims are inherited channel-workflow evidence, not revalidated by the daily scan.

  • Architecture With Skills update (2026-10-02): insert the supplied eShop prompts before the final Links slide (1555555247173140560). Channel workflow records v1.2.1 removing “and the 32px grid” from visible step 02 while retaining original prompts in notes; the later human message 1555558619376918659 says only “Drop this” and does not identify its target in text. Four-slide delivery and validation are recorded by that workflow, not rebuilt by this sync.

  • TrueNAS SCALE 26.0.0-BETA.2 at lego-cloud.eu ✅

  • Gitea: gitea.lego-cloud.eu (org: home-v1) ✅

  • General Gitea repository skill: ✅ gitea-repository-operations — reusable API + SSH workflow for any repo; full wiki runbook at wiki/infrastructure/gitea-repository-operations.md

  • Gitea API credential (confirmed 2026-08-02): use Bitwarden-provided HL_V1_GITEA_ACCESS_TOKEN for gitea.lego-cloud.eu; never use GITHUB_TOKEN for that instance or copy the Gitea token into /opt/data/.env. Bitwarden keys are exported without aliases and gateway restart is required after key changes.

  • Gitea recovery (confirmed 2026-08-11): Gitea 1.27.0 crashed all SSH/Smart-HTTP Git-serving processes despite healthy web/API; upgrade to 1.27.1 restored API and both Git transports. The queued VSSA commit 6fc42ab was pushed and remotely verified.

  • Gitea Pages pattern (verified 2026-08-11): Actions runners build repositories and publish to /gondor-v1-gitea-pages/<organization>/<repository>/, exposed at https://pages.apps.lego-cloud.eu/<organization>/<repository>/. The Pages dataset must be allowed by container.valid_volumes and mounted into job containers. skic-v1-playground/documentation run 2326 and the live site were verified.

  • Actions Docker (verified 2026-08-11): use the runner-mounted /var/run/docker.sock directly (client 29.6.2-1, Engine 29.0.4). DinD is unnecessary and privileged DinD is not currently supported by this runner.

  • CogArch GitHub attribution (required 2026-08-25): cogarch-github commits/PRs must disclose that Hermes acted using Oleg Lukasonok's PAT without implying Oleg performed or approved the action. When evidence permits, also name the Discord user who guided/originated the work; never invent uncertain attribution.

  • CogArch presentation governance (required 2026-09-04): channel-dedicated cogarch-storm is maintained in cognitive-architect-v1-skills-code-agent/cogarch-storm. Any IBM-flavored, IBM-originated, IBM-audience, or IBM-branded presentation must additionally use powerpoint-ibm-template; that skill must always reference/load powerpoint and powerpoint-advanced-python.

  • CogArch Solution Factory / Argo evidence boundary (verified 2026-09-15): exact live and archived label queries for architecture-id=arch_kiSj9jBKI returned no Workflow. The deployed Processing template contains conversion, chunking, image description, and merge steps, but completed Workflows are deleted after 24 hours and the accessible archive is empty. A separate arch_xOQ5jks0b High-Level Insights failure reached the code agent and exhausted retries with exit code 201; its overwritten detailed log reportedly showed operation timeout/retry-limit failure. A fail-open Redis DNS warning appeared in that failed run but did not recur in the read-only-verified rerun pod, so it is not the established cause.

  • CogArch Solution Factory BYOT MVP (decided 2026-09-16): use one persistent OpenShift Secret per user, created/patched/deleted by the existing settings backend; do not introduce IBM Secrets Manager or a new secrets service initially. Require server-derived opaque names, backend ServiceAccount permissions limited to Secret create/patch/delete without reads, admission restriction to labelled byot-* objects, trusted WorkflowTemplates, no token leakage, and fail-closed behavior. Running workflows may retain old values; provider revocation and cluster-loss recovery remain separate concerns.

  • Emplonet v1 channel setup (confirmed 2026-09-08): #emplonet-v1-general and #emplonet-v1-ingrida each have a dedicated local artifact workspace under /opt/data/channels/, a dedicated channel skill, and a private source repository in emplonet-v1-skills-code-agent; both are now included in the authoritative daily wiki scan and short config routing.

  • CogArch–SAP collaboration workspace (confirmed 2026-09-25): #cogarch-v1-collab-sap has a dedicated local workspace and private channel skill. Its private reference-v1-sap corpus covers 124 SAP Architecture Center source pages across 34 reference-architecture families and preserves 159 upstream Draw.io files with checksums and provenance; the first architecture family to assess remains open.

  • Lego CV workspace (confirmed 2026-09-14): #lego-v1-cv uses /opt/data/workspaces/channels/lego-v1-cv and the private lego-v1-skills-code-agent/lego-v1-cv channel skill. The supplied CV source was converted into brand-neutral editable DOCX and PDF artifacts without embedded IBM logos while retaining professional content.

  • Wiki health check (repaired 2026-09-13): /opt/data/scripts/health-check.sh no longer creates invalid multiline 0\n0 counts when grep -c finds no match, and its Discord test recognizes the gateway's actual inbound/send activity log forms. bash -n and a complete rerun passed all four checks; authenticated Discord REST still remains the source of truth for channel-scan coverage.

  • Wiki usability/format questions (open 2026-09-19): RootAtSkic wants a practical way to discover/retrieve channel wiki Markdown files instead of relying on numeric Discord-ID filenames, and asked whether migration to a Google “OKF” format would help. “OKF” and the desired delivery interface remain unclear; no migration is authorized.

  • CogArch Shorts presentation workflow (confirmed 2026-09-26): Architecture in an AI World is maintained in the dedicated Shorts workspace. Always return the latest deck as one image attachment per slide, in order; do not use a combined preview. The revised deck consolidates former slides 3–5 and 6–7, includes linked FINOS CALM, Structurizr, and ArchiMate reference slides, preserves design as a hard problem in the opening thesis, and removes the referenced “prompt enablement backbone” and “minutes” content (exact slide location was not stated in the source message).

  • CogArch Shorts current topic (decided 2026-10-01): set aside Architecture in an AI World without deleting it and return to the three-slide Shorts format for Architecture With Skills. Approved agenda order: cognitive-architect, cognitive-architect-2-archify, cognitive-architect-2-calm-finos, cognitive-architect-2-structurizr, cognitive-architect-2-archimate, each linking to its matching https://github.ibm.com/CTOTools-skills-code-agent/<name> repository. Evidence: 1555257448691011585, 1555257919635853365, 1555269400460787813; deck v1.1.0 is recorded by the channel workflow, not independently rebuilt by this sync.

  • Jev evaluation (open 2026-09-19): investigate Jev’s practical Home v1 uses, TrueNAS app availability/deployability, and likely impact; the request did not authorize deployment.

  • Operator communication preference (confirmed 2026-09-19): use retained knowledge of RootAtSkic’s character/principles and provide concise, decision-oriented summaries rather than large reading dumps.

  • Hermes delegation baseline (live 2026-09-12): max_concurrent_children: 10 and max_spawn_depth: 1; RootAtSkic asked how to increase sub-agent capacity, but no configuration change was authorized or made.

  • Home v1 skill-repository descriptions (status 2026-09-09): all 60 repositories in home-v1-skills-code-agent have non-empty Gitea descriptions, but the requested 3–5 sentence standard is met by 58/60. hermes-codex and development-durable-wave-execution still have one-sentence descriptions and remain open remediation.

  • TrueNAS channel skill (confirmed 2026-08-26): home-v1--truenas is maintained in private Gitea repository home-v1-skills-code-agent/home-v1--truenas and loaded directly from its checkout. Commit 634e0e6 passed 10 offline tests, Gitea Actions task 1398, remote readback, and a real authenticated read-only WebSocket query. Runtime access uses Bitwarden keys HL_V1_TRUENAS_URL and HL_V1_TRUENAS_API_KEY. The current IP-based URL fails strict certificate identity validation; use a certificate-matching DNS/CA path rather than permanent --insecure mode.

  • Gondor v1 (2026-08-02): Hermes SSH access as lego is verified to Osgiliath 000 and both Minas Tirith workers; use the gondor-v1-nodes skill. Open ingress incident: Argo CD's hostPort 80/443 ingress-nginx conflicts with Osgiliath's system Nginx; repair through home-v1/gondor-v1.git, not a transient live patch.

  • Hermes outage (2026-08-04–11): confirmed Codex quota exhaustion first blocked model responses while Discord/cron remained alive; a separate likely host/container-runtime failure then stopped all gateway, scheduler, log, and health progress after 2026-08-04 14:00:42 UTC. Physical restart at 2026-08-11 00:11:40 restored service without Hermes repair. Exact host-level trigger remains UNKNOWN because previous-boot TrueNAS/kernel/ZFS/runtime/hardware telemetry is unavailable inside the container.

  • world-v1/network-v1: ✅ global Internet address book; IP-level mapping now separates registry holder, route origin, operator, and service. Initial records: 0.0.0.0 and 1.1.1.1; wiki page at wiki/infrastructure/network-v1.md

  • Keycloak: keycloak.lego-cloud.eu ✅

  • Hermes Dashboard: jarvis-jr-v1.lego-cloud.eu ✅

  • STT Whisper: 192.168.1.218:39000 🔲

  • Piper TTS: 192.168.1.218:39001 🔲

Hermes Migration

  • Status: ✅ Live, cron active, backlog ownership officially handed to Hermes role (2026-07-16)
  • UPGRADED: Hermes is now sole agent on SKIC Discord server (2026-07-17 07:44 UTC, confirmed by Lego)
  • Hermes role ID: 1527260007307935776 | Discord permissions: Administrator ✅
  • morning-standup cron ✅ first successful fire: 2026-07-17 05:01 UTC; enhanced 2026-07-27 to include drawio-desktop#2499 monitoring (Lego directive: merge into standup, not separate job)
  • 🆕 Lego directive (2026-07-27 10:57 UTC): start building a dedicated skill for Hermes itself
  • Backlog files: wiki/vssa-dpvp/action-plan.md, wiki/cognitive-architect/action-plan.md
  • MartynasP model-access documentation: ⏳ still pending
  • Gitea wiki sync ✅ — pushing to skic-v1-playground/knowledge-wiki on every run (started 2026-07-17 09:30 UTC)
  • ✅ Gitea push restored — after Lego requested another access check at 11:08 UTC, SSH access to origin/main was verified at 15:08 UTC. Remote and local main both pointed to 35eb834; all previously queued commits are now on Gitea.
  • 🔴 Lego directive (2026-07-20 08:49 UTC): "please fix your issues yourself" — Hermes must resolve the Gitea push issue autonomously
  • ⚠️ Responsiveness incident (2026-07-20 09:54–11:03 UTC): Lego repeatedly asked whether the issues were fixed and what Hermes was doing; visible progress/status updates are expected during incident resolution.
  • Voice behavior (confirmed 2026-07-31): generate clear, human-friendly, conversational responses and let the participant guide with questions; avoid dense written-style delivery in live voice sessions.

Wiki Files

  • /opt/data/wiki/vssa-dpvp/action-plan.md — DPVP backlog
  • /opt/data/wiki/vssa-enterprise-architect/action-plan.md — EA consultancy
  • /opt/data/wiki/cognitive-architect/action-plan.md — CogArch backlog + pentest
  • /opt/data/wiki/sap-delivery/action-plan.md — SAP track
  • /opt/data/wiki/distinguished-engineer/action-plan.md — DE track
  • /opt/data/wiki/infrastructure/README.md — Services & endpoints
  • /opt/data/wiki/infrastructure/gitea-repository-operations.md — reusable Gitea repository runbook and skill reference
  • /opt/data/wiki/infrastructure/network-v1.md — global Internet address book and IP-level ownership/operation model
  • /opt/data/wiki/discord/README.md — Discord config
  • /opt/data/wiki/hermes-setup/README.md — Migration status
  • /opt/data/wiki/behavior/README.md — Operating rules