dw-kit 1.9.0-rc.1 → 1.9.1

package/.dw/core/PILLARS.md

@@ -1,122 +0,0 @@
-# dw-kit v2.0 — 5 Pillar Architecture
-
-dw-kit v2.0 positions itself as a **Context-First SDLC Governance Layer** — not a prescriptive workflow engine. AI drives execution; dw-kit provides guardrails, context, and decision trail.
-
-**Framing inversion:** `prescriptive workflow → descriptive governance`. Moat is organizational memory compounding over time — something IDE tools (Cursor, Copilot) structurally cannot own because they are session-scoped.
-
----
-
-## Pillar 1: GUARDS — Block unsafe actions
-
-**Role:** Non-negotiable safety boundaries. Enforced by hooks, zero discretion.
-
-**Components:**
-- `.claude/hooks/privacy-block.sh` — Prevent reading `.env*`, `credentials*`, `*.pem`, key files
-- `.claude/hooks/pre-commit-gate.sh` — Quality checks + sensitive data scan before commits
-
-**Obsolescence test:** AI gets smarter → safety becomes MORE important (velocity × risk). These hooks never obsolete.
-
-**Team impact:** Prevents accidents at individual dev level, reduces incident count team-wide.
-
----
-
-## Pillar 2: SURFACES — Make state visible
-
-**Role:** Shared team context. Human and AI both read.
-
-**Components:**
-- `.dw/tasks/ACTIVE.md` — Auto-generated index of active tasks (TechLead cat to see team state)
-- `.dw/context/project-map.md` — Module structure and boundaries
-- `.dw/context/modules/*.md` — Per-module documentation
-- `CLAUDE.md` + `.claude/rules/dw.md` — Auto-injected conventions
-
-**Obsolescence test:** AI gets smarter → teams coordinate more ambitiously → surfaces become MORE load-bearing.
-
-**Team impact:** New dev onboards from surfaces alone. TechLead audits without asking devs "what are you doing?"
-
----
-
-## Pillar 3: RECORDS — Capture decisions
-
-**Role:** Organizational memory. The WHY behind architectural choices.
-
-**Components:**
-- `.dw/decisions/{NNNN}-{title}.md` — ADRs with structured YAML header
-- `/dw:decision` skill — Interactive ADR wizard
-- Status tracking: `Proposed | Accepted | Deprecated | Superseded by ADR-{NNNN}`
-
-**Obsolescence test:** AI gets smarter → needs WHY context to avoid technically-correct but strategically-wrong decisions. ADRs become MORE valuable.
-
-**Team impact:** Replaces scattered Slack threads. 6-month-old devs find decision trail. New hires understand architecture fast.
-
-**Unique moat:** IDE tools (Cursor, Copilot) structurally can't own this — they're session-scoped, ADRs are cross-session artifacts.
-
----
-
-## Pillar 4: BRIDGES — Connect across sessions
-
-**Role:** Continuity over time. Handle the "long chat → no handoff → team lost context" problem.
-
-**Components:**
-- `.dw/tasks/{task}/tracking.md` — Mutable progress log with friction journal
-- Stop hook auto-handoff — Appends session summary to tracking.md on uncommitted changes
-- (v2.0+) Living docs detection — Flag when code diverges from docs
-
-**Obsolescence test:** AI sessions remain ephemeral regardless of capability. Bridges always needed.
-
-**Team impact:** Pick-up-where-left-off works across devs and across sessions. TechLead sees real velocity from actual logs, not status meeting summaries.
-
----
-
-## Pillar 5: TUNES — Behavioral knobs
-
-**Role:** Team/solo customization. Governs how Pillars 1-4 behave.
-
-**Components:**
-- `.dw/config/dw.config.yml` — Master config (depth, roles, flags)
-- `.dw/config/dw.config.local.yml` — Machine-specific override (gitignored)
-- Presets: `solo` · `team` · `enterprise`
-- Depth routing: `quick` · `standard` · `thorough`
-- Role system: `dev` · `techlead` · `ba` · `qc` · `pm`
-
-**Obsolescence test:** Team size and preferences always vary. Config always needed.
-
-**Team impact:** Same dw-kit serves vibe coder (preset solo) and enterprise team (preset enterprise) with same core.
-
----
-
-## Pillar Cross-References
-
-| Question | Answer location |
-|----------|----------------|
-| "Is this safe to do?" | Pillar 1 (Guards) — hooks block |
-| "What's everyone working on?" | Pillar 2 (Surfaces) — ACTIVE.md |
-| "Why did we choose X?" | Pillar 3 (Records) — ADRs |
-| "Where were we yesterday?" | Pillar 4 (Bridges) — tracking.md |
-| "How does our team work?" | Pillar 5 (Tunes) — config + presets |
-
-## Design Principles
-
-1. **Descriptive, not prescriptive** — AI chooses approach; dw-kit supplies context + safety
-2. **Obsolescence-aware** — Every feature passes "more valuable if AI smarter?" test
-3. **Dual audience** — Solo + team from same core, different defaults
-4. **Data-driven evolution** — Telemetry guides cut decisions, not gut-feel
-5. **Escape hatches** — `--no-dw`, `legacy_features: true`, `DW_NO_TELEMETRY=1`
-
-## Future: Pillar 6 — JANITORS (deferred to post-v2.0)
-
-**Status:** Draft, deferred — see `.dw/decisions/0003-pillar-6-janitors.md`
-
-**Role:** Reactive cleanup of AI-generated waste. Current 5 pillars are *preventive* (govern what goes in); Janitors governs *what stays*.
-
-**Rationale:** When 99% of code is AI-generated, prevention alone cannot scale. Inspired by urban waste management — cities use multi-tier systems (sort → collect → recycle → regulate), not just "don't litter."
-
-**Revisit:** After v2.0 GA (post 2026-08-15), based on real-world friction data.
-
-## Versioning
-
-- **v1.3** — Foundation (new format, scaffolding, telemetry)
-- **v1.4** — Data-driven cuts based on telemetry evidence
-- **v2.0** — Unified release with full 5-pillar integration
-- **v2.1+** — Advanced features (auto-handoff LLM, cross-repo, dashboard UI)
-- **v2.2+** — Janitors pillar (if validated by v2.0 friction data)

package/CLAUDE.md

@@ -1,44 +0,0 @@
-# dw-kit (repo)
-
-Workflow toolkit codebase. Rules live in `.claude/rules/` (auto-loaded).
-
-**v2.0 direction:** Context-First SDLC Governance Layer (5 pillars — see `.dw/core/PILLARS.md`)
-**Current:** **v1.8.0-rc.1** (2026-05-25) — 4/5 phase substrate for [G-rgoal-realtime-orch](.dw/goals/G-rgoal-realtime-orch/goal.md) voice-meeting Root Goal: persistent CLI-agent session runtime (`dw session *`), Telegram chat bridge (`dw connector telegram setup`) with 1-command interactive wizard, multi-workspace registry (`dw workspace *`), and browser voice MVP (`dw voice`) with hybrid orchestrator fallback (Claude/Codex/Gemini), full bilingual UX (en + vi), voice-not-installed fallback via Google Translate TTS proxy. **v1.7.0 stable** on `main` (2026-05-24). 368/368 smoke tests. Active ADRs: ADR-0001 (Pragmatic Lean), ADR-0005/0006 (Supply-Chain Guard; sunset review 2026-08-12), ADR-0008 (Task Docs v3, v1.5), ADR-0009 (Agent OS, v1.6), ADR-0010 (Goals Layer, v1.7), ADR-0011 (Session Runtime + Voice Orchestrator, v1.8 — Accepted), ADR-0014 (Orchestrator Action Execution — Accepted), ADR-0015 (Multi-Agent Live Debate — Accepted), ADR-0016 (`dw:goal` Outcome-Driven Pursuit Loop — Accepted).
-
----
-
-## Tech Stack
-
-- Runtime: Node.js ≥18, ESM (`.mjs`)
-- CLI: `commander` · UI: `enquirer`, `chalk` · Config: `js-yaml`, `ajv`
-- Tests: `node src/smoke-test.mjs`
-
-## Repo Structure
-
-```
-bin/          CLI entrypoint
-src/
-  commands/   CLI subcommands (init, upgrade, dashboard, metrics, ...)
-  lib/        Shared utilities (config, telemetry, active-index, ...)
-.claude/
-  hooks/      Bash hooks (Guards pillar)
-  rules/      dw.md (consolidated) + code-style + commit-standards
-  skills/     Slash commands with dw:* namespace
-.dw/
-  core/       WORKFLOW · THINKING · QUALITY · ROLES · PILLARS · templates/
-  decisions/  ADRs (Records pillar)
-  tasks/      Active + archive/ (Bridges pillar — via task.md v3 / tracking.md v2)
-  metrics/    Local telemetry (events.jsonl)
-  config/     dw.config.yml
-  security/   IoC namespace fixture (Guards pillar — ADR-0005)
-  research/   Investigation notes, RFC-style proposals, voter panel outputs
-```
-
-## Dev Notes
-
-- All source ESM — no CommonJS
-- `TOOLKIT_ROOT` resolved from `import.meta.url` in each command
-- Hooks python3-free (node only — Windows compat)
-- `dw-kit-evolve` + `dw-kit-audit` are maintainer-only — excluded from npm package
-- Published package files declared explicitly in `package.json#files`
-- Telemetry local-only, `DW_NO_TELEMETRY=1` to disable