Co-authored-by: Jeffrey Smith <jasafpro@gmail.com> Co-committed-by: Jeffrey Smith <jasafpro@gmail.com>
7.0 KiB
ROADMAP — v0.10.x+ Version Shift (Final)
Summary
Four new series inserted after v0.9.x. Panels (kernel primitive), Notes
(first reference extension), Chat (second reference extension), and the
remaining reference libraries/extensions including llm-bridge with
the tool meta-tool pattern.
Full Roadmap
| Series | Title | Versions | Design Doc |
|---|---|---|---|
| v0.9.x | Multi-Surface + Workflow Redesign | 10 | — |
| v0.10.x | Panels + Composable Layout | 5 | DESIGN-panels.md |
| v0.11.x | Notes Reference Extension | 11 | DESIGN-notes-v011x.md |
| v0.12.x | Chat Reference Extension | 9 | DESIGN-chat-v012x.md |
| v0.13.x | Reference Libraries + Extensions | 9 | — |
| v0.14.x | Sidecar Tier + Polish | 3 | — |
| v1.0.0 | Stable Release | — | — |
v0.10.x — Panels + Composable Layout
| Version | Title |
|---|---|
| v0.10.0 | Panel Manifest + Lifecycle |
| v0.10.1 | FloatingPanel Primitive |
| v0.10.2 | Docked Panels + Mode Transitions |
| v0.10.3 | Panel Communication Patterns |
| v0.10.4 | Reference Panel: Notes (basic) |
v0.11.x — Notes Reference Extension
| Version | Title |
|---|---|
| v0.11.0 | UI/UX Foundation |
| v0.11.1 | Deep Folders + Navigation |
| v0.11.2 | Wikilinks + Backlinks |
| v0.11.3 | Live Preview + Rich Editing |
| v0.11.4 | Note Sharing + Permissions |
| v0.11.5 | Graph + Outline Hardening |
| v0.11.6 | Quick Switcher + Commands |
| v0.11.7 | Daily Notes + Templates |
| v0.11.8 | Transclusion + Embeds |
| v0.11.9 | Composability: Slots + Actions |
| v0.11.10 | Panel Enhancement + Quality Gate |
v0.12.x — Chat Reference Extension
Human-to-human first. AI via llm-bridge (v0.13.x).
| Version | Title |
|---|---|
| v0.12.0 | UI/UX Foundation |
| v0.12.1 | Conversation Folders + Attributes |
| v0.12.2 | Reactions + Threads + Pins |
| v0.12.3 | Rich Compose + Attachments |
| v0.12.4 | @Mentions + Notifications |
| v0.12.5 | Link Previews + Message Formatting |
| v0.12.6 | Conversation Themes + Personality |
| v0.12.7 | Composability: Slots + Actions |
| v0.12.8 | Panels + Quality Gate |
v0.13.x — Reference Libraries + Extensions
Core libraries and remaining extensions. llm-bridge is the key
delivery — extends both notes and chat through composability primitives,
and introduces the tool meta-tool pattern.
| Version | Title |
|---|---|
| v0.13.0 | vector-store Library |
| v0.13.1 | llm-bridge Core Library |
| v0.13.2 | llm-bridge → Chat: Multi-Persona Context + Tool Meta-Tool |
| v0.13.3 | llm-bridge → Notes: AI Toolbar Actions |
| v0.13.4 | file-share Extension |
| v0.13.5 | code-workspace Extension |
| v0.13.6 | image-gen + image-edit Extensions |
| v0.13.7 | Tool Meta-Tool Hardening + Scoping |
| v0.13.8 | Integration Quality Gate |
v0.13.1 — llm-bridge Core:
Model abstraction, provider BYOK via connections, complete(),
embed(), classify(). Persona CRUD (name, avatar, system_prompt,
visibility). Layered prompt architecture (6 layers: admin safety →
folder context → persona identity → conversation → tools → user).
Admin safety rails as non-overridable platform setting.
v0.13.2 — Chat Integration:
Contributes folder attributes (system_prompt, context_policy,
model_override, allowed_tools) to chat. Contributes "Invite AI" /
"Dismiss AI" to chat:participant-actions. Contributes "AI Reply" /
"AI Summarize" to chat:message-actions. Contributes "Ask AI" to
chat:composer-tools. Context archetype implementation (resident /
scoped / stateless). Session tracking, gap handling, memory extraction
on dismiss. Tool meta-tool v1: sw.actions.list() → LLM tool
definitions. AI calls extension actions, results flow back into chat.
v0.13.3 — Notes Integration:
Contributes "AI Summarize" / "AI Translate" / "AI Fix Grammar" to
notes:toolbar-actions. Contributes /ai to notes:slash-commands.
Uses notes.get / notes.search actions for context.
v0.13.7 — Tool Meta-Tool Hardening:
Three-tier tool scoping: admin blocklist (global), folder allowed_tools
(space-level), persona allow/deny (character-level). Tool result
rendering through sw.renderers. Rate limiting on tool calls. Audit
logging of tool use. Error handling (tool failure → graceful message).
v0.13.8 — Integration Quality Gate: AI participant lifecycle tested end-to-end. Tool meta-tool demonstrated with 3+ extensions. All three context archetypes tested. Persona creation/sharing across visibility levels. Admin safety rails validated (prompt injection resistance). Performance: completion latency measured.
v0.14.x — Sidecar Tier + Polish
| Version | Title |
|---|---|
| v0.14.0 | Sidecar Tier |
| v0.14.1 | Native Dialog Audit |
| v0.14.2 | Stability + Migration Tooling |
v1.0.0 Gate Criteria
- Notes reference extension shipped (all v0.11.x)
- Chat reference extension shipped (all v0.12.x)
llm-bridgeextends both notes and chat through composability- Tool meta-tool demonstrated: AI uses 3+ extension actions in a single conversation turn
- Notes and chat UI reviewed against design principles
- At least 2 panels consumed cross-package
- At least 2 slot contributions per host surface demonstrated
- Multi-persona context archetypes demonstrated
- Admin safety rails validated
- Note and conversation sharing functional end-to-end
- Headless E2E green on PG + SQLite
- All kernel Starlark modules documented
- All API routes covered by OpenAPI spec
- Upgrade path tested from v0.8.0 → v1.0.0
- Single-binary + Docker + K8s deployment paths documented
Design Decisions Log
| Decision | Rationale |
|---|---|
| Panels as kernel primitive (v0.10.x) | Z-index coordination, drag/resize, layout negotiation are kernel concerns. |
| UI/UX redesign as first version in each reference series | Every feature builds on the visual foundation. |
| Notes before chat | Notes is simpler (no realtime) and proves storage/rendering/composability. Chat adds realtime + llm-bridge story. |
| Chat human-to-human first | AI is an extension concern. Keeps chat testable and usable standalone. |
| Folder attributes as extension bridge | Chat stores attributes it doesn't understand. llm-bridge contributes definitions. Zero coupling. |
| Action registry as tool registry (meta-tool) | Dynamic, zero-config AI tool-use. Installed extensions = AI capabilities. Uniquely Armature. |
| Layered prompt architecture (6 layers) | Admin safety not overridable. Clear separation: platform → space → character → context → tools → user. |
| Personas in llm-bridge, not chat | Chat sees AI as just another participant_type. Persona identity is llm-bridge's concern. |
| Background images via CSS filter | No server-side processing. filter: blur() brightness() saturate() + opacity handles any image. |
| Soft panel deps for notes↔chat | Runtime-resolved, no circular dependency. Both function independently. |
| llm-bridge after notes + chat | Proves the composability hooks work without being designed for a specific consumer. |