claude-projects/claude-config/research/INDEX.md

Research Index

Catalog of research on how we use Claude / Claude Code and where we want to take it. One row per topic. Add a row whenever a new research file lands here. Findings in this space move fast — re-verify versions, repos, and Linux support before acting on anything.

Topic	File	Status	Key open decision
Infrastructure synthesis — the whole landscape + reconciliation of the new research against decisions already made	infrastructure-synthesis.md	Synthesized 2026-09-08	Confirm the July→Sept reality (Max upgrade / Voice-Chat / Tailscale-Twingate / Agent-Sudo) before acting on anything.
Autonomy + isolation evaluation — does the plan work, gaps, and the 2026 Docker/Anthropic isolation landscape (auto mode, Bash sandbox, sandbox-runtime, Docker Sandboxes microVMs)	autonomy-isolation-evaluation.md	Evaluated 2026-09-08 (Fable 5.1)	DECIDED 2026-09-08 (adopt, autonomous runner first) → decisions/GAMEPLAN_security-infra-deploy.md. 16:07 outage = boot order (LAN-IP port binds before WiFi), solved.
Local AI coding stack (inference engine, coding harness, LifeOS, voice, skill porting)	local-ai-coding-stack-research.md	Surface-level, in progress	Goal framing: RESOLVED by the existing vision = cost-reduction + tooling-independence, Claude stays the brain (NOT fully-local). See synthesis Part 3.
AI harness layer — scored comparison of the layer around Claude Code (LifeOS/PAI, SuperClaude, Claude-Flow, oh-my-claudecode, claude-mem, Basic Memory, Letta) vs baseline under R1 Pro-OAuth / R2 merge / R3 FOSS	ai-harness-layer-research.md	Scored 2026-09-08 (Fable 5.1)	Baseline wins (14/21). Fix /recall embed endpoint first; only then consider a sandboxed Basic Memory trial (11/21). Claude-Flow + claude-mem DQ on R1 (Anthropic legal page: Agent SDK / executors need API keys).
jcode vs Claude Code — tests the two jcode hopes (bigger context; "CLAUDE.md size limit") and ranks executors (Claude Code, jcode, OpenCode, Aider, Goose, Cline, OpenHands, Gemini CLI) under R1 Pro-OAuth / R2 hooks+skills / R3 FOSS	jcode-vs-claude-code.md	Scored 2026-09-08 (Fable 5.1)	Claude Code stays (14/18). Every OSS executor DQ on R1 for Claude (jcode spoofs the Claude Code client; OpenCode removed Pro/Max under legal request). "Bigger window" = model property: Sonnet 5 is 1M free on Pro, Opus 1M = paid credits. Pain (b) is the 25 KB MEMORY.md index cap, not CLAUDE.md. Gemini CLI = optional non-Claude secondary for long reads.
HyperAgent marketing fit — disambiguation (Airtable Hyperagent vs Hyperbrowser OSS vs FSoft paper), constraints, per-job fit for API-clients / KDP / Etsy, ToS risk, OSS alternatives	hyperagent-marketing-fit.md	Assessed 2026-09-08 (Fable 5.1)	Reject all three jobs. User likely heard of closed-source hyperagent.com (credit-metered, $4–35/task). OSS HyperAgent has baseURL for Ollama but hard-codes headless:false; KDP/Etsy/LinkedIn browser automation = ban risk. Only fragment worth anything = plain Playwright enrichment under N8N; browser-use beats it if a trial is ever wanted.

| Omarchy OS evaluation — is DHH's Omarchy (Arch+Hyprland) a fit as a laptop daily driver (UC1) and/or as server-01's OS (UC2) vs the planned plain-Debian migration; "AI features like Claude skills" truth-check | omarchy-os-evaluation.md | Researched 2026-09-13 (Opus 4.8), preliminary/hedged | UC1 laptop = lean-yes/depends-on-GPU (MED); UC2 server = lean-no (HIGH). "Claude skills" is a conflation: Omarchy pre-wires the normal Claude Code CLI (works on Pro OAuth, no API key) + ONE experimental system-tailoring skill — no bundled model. Main session: infra-fit synthesis + verify laptop GPU/disk/partition + snapshot-rollback caveat (#8047) before any install. | | Omarchy AI skills deep-dive — full skill inventory + the crash auto-investigation skill (does it exist? mechanism?) + portability to our Debian/LMDE machines + a blueprint mapping it onto our diagnose skill / Hermes / NTFY / sudo-bridge | omarchy-ai-skills-deep-dive.md | Researched 2026-09-13 (Opus 4.8), preliminary/hedged | Crash skill is REAL & shipped (diagnose-crash, systemd-coredump journald watcher → click-to-diagnose), but it is diagnose-and-report ONLY — explicitly does NOT auto-fix ("Diagnosis reads; it does not fix") and is human-click-triggered, not unattended → the "then fixes it" the user saw is NOT the shipped skill. Two skills ship (diagnose-crash + omarchy tailoring). Portability = mixed: method lifts nearly verbatim, but Arch (pacman/archlinux-debuginfod) + Hyprland/Quickshell trigger chain need rewrite for headless Debian. Main session: decide whether to spec+build a ported diagnose-and-report skill (fixes stay behind sudo-bridge). | | GitHub policy vs lead pipeline — may the GitHub-sourced cold-email outreach resume? | github-aup-lead-pipeline.md | Verdict 2026-09-08 | NO — AUP §7 + ToS §H forbid unsolicited email from API-collected data (verbatim quotes in file). Sourcing + contact move to the marketing overhaul (business_projects 43). | | KDP business evaluation (Prompt E) — J.M. Hartley AI-assisted Q&A books: KDP 2026 rules (AI disclosure, low-content, royalty/print cost), 9-niche scoring, $0 marketing plan, ads model, kill criteria | kdp-business-evaluation.md | Evaluated 2026-09-09 (Fable 5.1) | GO WITH CHANGES. Drop the generic "100,000 Whys" clone (a #138-in-Books colour incumbent exists, 2026-05); 3 shortlisted niches (seniors large-print Q&A 16/18, single-theme kids why 15/18, couples questions 14/18). First sale at $0 realistic in 21–60 days (LOW-MED); royalties pay ~60 d after month end → ads ($90/mo) only from ~Dec 2026. Claude text = AI-generated → must disclose. | | Etsy / ConfettiPrintCo evaluation — AI-designed POD (Printify) + printables viability, official Etsy rules/fees (AI disclosure, production partner, digital limits), $0-ad marketing plan, unit economics, kill criteria (business_projects 46) | etsy-business-evaluation.md | Evaluated 2026-09-09 (Fable 5.1, Prompt F) | go_with_changes — digital-first (junk-journal/planner printables), Printify POD only after a digital bestseller badge + $50 profit. First sale at $0: 21–56 days after 30 listings, ~55%. Opt OUT of Offsite Ads day 1. Demand counts UNVERIFIED (etsy.com 403) → owner runs browser method before 2026-09-14. | | Mesh-VPN / zero-trust access — NetBird vs Tailscale — final defensible pick for connecting primary host + server-01 (sandbox) + roaming laptop/phone into one private mesh, weighted by self-host-first + cost + zero-trust + solo-operator burden | netbird-vs-tailscale-mesh-vpn.md | Decided 2026-09-15 (Opus 4.8) | NetBird (self-hosted control plane). Wins on principle #1 (own the coordination plane; Tailscale's is proprietary SaaS, self-host = 3rd-party Headscale w/ gaps). Weighted 3.98 vs TS-cloud 3.19. Flip → Tailscale-cloud only if owner accepts vendor lock-in on this one tool to save solo-operator upkeep. Open: Authelia-vs-Authentik OIDC (P4), TURN-through-cloudflared, mgmt DB engine — all UNVERIFIED, test on server-01 first. | | NetBird + Twingate deployment gameplan — owner-decided plan to run BOTH tools with distinct roles (Twingate=partner access to Nextcloud over Tyler's CGNAT; NetBird=owner device mesh); phases 0 preflight / 1 Twingate+partner-swap / 2 NetBird self-host; Austin offboard + Hailee onboard in scope | netbird-twingate-gameplan.md | Planned 2026-09-15 (Opus 4.8) | Bootstrap-vs-self-host for NetBird STILL OPEN (preflight resolves DB engine / Authelia-OIDC / TURN-through-cloudflared to inform it). Twingate first (deadline Sun 09-20). server-01 = validation sandbox only; all real services on primary server. Prompts K (preflight) + L (deploy). | | API business (RapidAPI) evaluation + marketing plan (Prompt D) — can RapidAPI produce first paid subscribers at $0 ads; 8-category scoring; two API candidates; no-social/no-outreach marketing plan; N8N/local-model automation + training-log; kill criteria (business_projects 44) | api-business-evaluation.md | Evaluated 2026-09-09 (Fable 5.1, agent D) | go_with_changes — build a utility API (HTML/MD→PDF + PDF→JSON) not a second finance niche; first sale at $0 realistic but 60–120 days (LOW-MEDIUM) → slowest of the three, run in background. RapidAPI fee 25 % since 2025-11-15, PayPal-only; category pages login-walled (no public demand data); ToS forbids using marketplace info to solicit consumers → quota upsell = opt-in via own docs site. | | Digital-businesses marketing strategy (agent J) — the layer above D/E/F: launch order + bootstrap-funding chain, one reusable "Listing Engine" marketing system (project 43: keyword intake → listing draft/QA → product-embedded opt-in → weekly metrics/kill-rule, 5 unified marketing_* tables + one hub site), per-business launch-week checklists (what is NEW vs each plan), automation/training-log layer, weekly metrics | digital-businesses-marketing-strategy.md | Written 2026-09-15 (Fable 5.1, agent J) | Etsy → KDP → API locked (swap Etsy/KDP only if the browser demand method fails this week). All three have a $0 first-sale path (Etsy ~55 %, KDP/API LOW-MED). NEW: Etsy Share & Save day 1; KDP title A by 09-25 with bank/tax interview day 1; API = background ≤ 2 h/wk until 10-05. Six grill-me Qs incl. two scope conflicts (biz_projects 39 birthday printables vs junk-journal; 40 couples vs seniors). Tables/hub site PROPOSED, not built. | | Partner access audit — Austin offboard / Hailee onboard (preflight K, read-only) — every place departed partner Austin has access + every place new 33% partner Hailee must be added; offboard/onboard checklists | partner-access-audit-austin-hailee.md | Audited 2026-09-15 (Opus 4.8) | ZERO changes made. Austin access = Nextcloud user Austin_Mktg (enabled) + /Partner Meetings share id 2 (only that). No groups, no Vault refs. Bitwarden bridge DOWN → manual owner check. Talk rooms not occ-listable → disabling the user covers Talk (verify in UI). External (Discord/email) = manual. Hailee: new NC user + /Partner Meetings share + Talk room + Twingate Partners group. NOT Grim/Josh. | | Twingate deploy runbook (preflight K) — partner access to Nextcloud over Tyler's CGNAT; free-tier, connector container shape, Remote-Network/Resource/Group model, client requirement, Vault paths | twingate-deploy-runbook.md | Written 2026-09-15 (Opus 4.8) | Free tier fits (5 users/10 nets/50 resources; partners=Tyler+Hailee). Connector twingate/connector on PRIMARY server, outbound-only → CGNAT-proof, no port-forward. Client app REQUIRED (no clientless). Vault secret/twingate/* (none exist yet). UNVERIFIED: Talk WebRTC fully over Twingate + browser LNA → test before dropping Open Relay TURN. | | NetBird self-host buildplan + bootstrap decision aid (preflight K) — owner device mesh; resolves the 3 open items + bootstrap-vs-selfhost recommendation + server-01 sandbox runbook + primary-server prod shape | netbird-selfhost-buildplan.md | Written 2026-09-15 (Opus 4.8) | (a) DB: SQLite default, Postgres supported → use homelab Postgres. (b) OIDC: Authelia WORKS via generic-OIDC (official Authelia↔NetBird guide) — no Authentik; device-flow for CLI UNVERIFIED (setup-keys fallback). (c) TURN: coturn needs a UDP port; WSS relay on 443 may avoid it (UNVERIFIED — test on server-01). Recommend BOOTSTRAP on NetBird Cloud free tier first (deciding factor: 09-20 Twingate deadline owns this week). | | secrets-proxy S1 investigation (#191 → #174, read-only) — why it stopped 07-09, endpoint/auth inventory, 18 compose vars → Vault, callers (Jenkins promote), SP1–SP4 validity, corrected SP1 draft | secrets_proxy_S1_investigation.md | Investigated 2026-09-27 (Opus 5.5) | Do NOT restart as-is (bash -c gate bypass, docker.sock = host root, env leak). SP plan partly valid. Owner: pick promote path (/exec deploy_service recommended) before SP1; verify transit key secrets-proxy-proxymd (UNVERIFIED). | | Voice Chat TTS engine re-check (V0, #214) — XTTS-v2 (idiap) vs F5, Chatterbox, Fish S2, CosyVoice 3, IndexTTS 2.5, Zonos, Qwen3-TTS + Kokoro/Piper baselines under R1 local / R2 clone / R3 8 GB Turing / R4 stream / R5 maintained / R6 licence | /opt/appdata/docker/research/voice_tts_engine_2026-09-28.md | Researched 2026-09-28 (background agent, Opus 5.5) | Switch primary to Chatterbox-Turbo (MIT code+weights, devnen server /tts stream, fp32 on Turing) ; XTTS-v2 = fallback (CPML non-commercial, Coqui defunct). Changes July decision. V2 must smoke-test VRAM + first-audio on the 2060 SUPER. |

Where the prior infrastructure research lives (memory corpus)

Not duplicated here — cited in infrastructure-synthesis.md. Key files: project_two_horizon_operating_model (governing filter), project_ai_infrastructure_vision (Horizon-B master), reference_open_source_ai_tools, reference_available_tools, security_principles + feedback_autonomous_security_constrain_not_gate + playbook_zero_trust_security, project_jenkins_hermes_boundary, project_tailscale_twingate_deploy, project_voice_chat_claude_code, agent-builder/ (Agent-Sudo / Constrained Autonomy design + gameplan).

Cross-cutting open decisions

Pulled up from the individual briefs so they don't get buried:

0. ✅ DECIDED 2026-09-08 — Isolation boundary: Incus KVM VM on server-01, non-root, egress allowlist, brokers only. Decides whether --dangerously-skip-permissions/auto mode is safe. (autonomy-isolation-evaluation.md §4)

1. The goal — cost / independence / fully-local. Governs every other choice in the local stack. (from local-ai-coding-stack-research.md §0)

2. 8 GB VRAM ceiling — RTX 2060 Super caps a fully-local setup at ~7–8B @ Q4. Which coding models actually fit with agentic context, and how reliable is their tool-calling through a harness? (§1, §2)

3. Daily-driver harness — likely OpenCode for stability; jcode only if its swarm/multi-agent feature earns it. Confirm canonical jcode repo/language before investing. (§3)

4. LifeOS — pin current version, re-check real Linux status per component (docs overstate it), confirm whether LifeOS↔harness pairing needs adapter work. (§4, §5)

5. Voice input — local Whisper dictation vs. reviving LifeOS VoiceServer (needs macOS→LMDE port). (§6)

To fold in

• Findings from the earlier Claude-chatbot research session that prompted creating this directory (the reason local-ai-coding-stack-research.md exists). Bring those notes in as their own file(s) and index them here.