safeword 0.2.4 → 0.2.6

package/.safeword/tickets/002-architecture-enforcement.md

@@ -1,211 +0,0 @@
----
-id: 002
-status: in_progress
----
-
-# Architecture Enforcement for SAFEWORD
-
-**Goal:** Add layers/boundaries enforcement to SAFEWORD so AI agents respect architectural boundaries.
-
-**Why:** Without guardrails, AI agents create circular dependencies, god modules, and unsafe refactoring patterns.
-
-## Work Log
-
-- 2025-11-27T00:00:00Z Started: Created plan at `.safeword/planning/005-architecture-enforcement-system.md`
-- 2025-11-27T00:00:00Z Decision: Extend `setup-linting.sh` for eslint-plugin-boundaries (not new script)
-- 2025-11-27T00:00:00Z Decision: Add layers/boundaries section to existing `architecture-guide.md`
-- 2025-11-27T00:00:00Z Decision: Create `prompts/` folder, extract existing quality-review prompt from `setup-quality.sh`
-- 2025-11-27T00:00:00Z Decision: Drop Biome support entirely, ESLint-only with eslint-plugin-boundaries
-- 2025-11-27 Implementation completed:
-  - Phase 1: Added Layers & Boundaries section to `architecture-guide.md`, created `templates/architecture-template.md`
-  - Phase 2: Updated `setup-linting.sh` - removed Biome mode, added eslint-plugin-boundaries to all ESLint modes
-  - Phase 3: Created `framework/prompts/` with quality-review.md and arch-review.md, created `scripts/arch-review.sh`, updated setup-safeword.sh and setup-quality.sh
-  - Phase 4: Created `templates/ci/architecture-check.yml`
-
----
-
-## Planning Docs
-
-- `.safeword/planning/005-architecture-enforcement-system.md` - Full plan with phases, deliverables, tool selection
-
-## User Stories
-
-**Source:** Stories 1-11 from `001-guides-review-user-stories.md` (documentation guidance). Stories 12-19 added for enforcement/tooling.
-
----
-
-### Story 12: Layer Definition Guidance ✅
-
-**As a** developer setting up a new project
-**I want** clear layer definitions (app, domain, infra, shared) with responsibilities
-**So that** I know how to organize code from the start
-
-**Acceptance Criteria:**
-
-- [x] Layer definitions table in `architecture-guide.md`
-- [x] Each layer has directory and responsibility defined
-- [x] `architecture-template.md` includes same structure
-
----
-
-### Story 13: Dependency Rules ✅
-
-**As a** developer writing cross-layer code
-**I want** an allowed dependencies matrix showing what can import what
-**So that** I avoid circular dependencies and leaky abstractions
-
-**Acceptance Criteria:**
-
-- [x] Matrix shows from/to/allowed/rationale
-- [x] Forbidden imports marked with ❌
-- [x] Rationale explains why each rule exists
-
----
-
-### Story 14: Edge Case Handling ✅
-
-**As a** developer with non-standard project structure
-**I want** documented edge cases (brownfield, monorepo, non-3-layer)
-**So that** I can adapt boundary rules to my situation
-
-**Acceptance Criteria:**
-
-- [x] Edge cases table in guide
-- [x] Solutions for each scenario
-- [x] Brownfield adoption path (warnings-only → incremental fix → enforce)
-
----
-
-### Story 15: Static Boundary Enforcement ✅
-
-**As a** developer
-**I want** boundary violations caught by ESLint
-**So that** I see errors immediately in IDE and CI
-
-**Acceptance Criteria:**
-
-- [x] `eslint-plugin-boundaries` added by `setup-linting.sh`
-- [x] Setup instructions + common issues in `architecture-guide.md`
-
----
-
-### Story 16: LLM Architecture Review ✅
-
-**As a** developer committing changes
-**I want** an LLM to review for architectural anti-patterns
-**So that** semantic issues are caught that static analysis misses
-
-**Acceptance Criteria:**
-
-- [x] `arch-review.sh` calls Haiku API
-- [x] Returns JSON verdict: `clean` | `minor` | `refactor_needed`
-- [x] Reviews against `ARCHITECTURE.md` if present
-- [x] Checks for: misplaced logic, god modules, leaky abstractions, tight coupling, boundary violations
-
----
-
-### Story 17: Pre-commit Enforcement ✅
-
-**As a** developer
-**I want** boundary violations to block my commit
-**So that** architectural issues are caught before code review
-
-**Acceptance Criteria:**
-
-- [x] Pre-commit hook runs eslint + arch-review.sh
-- [x] Lint errors block commit
-- [x] `refactor_needed` verdict blocks commit
-- [x] `minor` verdict warns but allows commit
-
----
-
-### Story 18: CI Architecture Check ✅
-
-**As a** team lead
-**I want** PRs checked for boundary violations in CI
-**So that** architectural issues can't slip through code review
-
-**Acceptance Criteria:**
-
-- [x] `architecture-check.yml` template provided
-- [x] Runs tsc, eslint (with boundaries), optional LLM review
-- [x] LLM review is non-blocking (advisory only)
-
----
-
-### Story 19: Architecture Template ✅
-
-**As a** developer starting a new project
-**I want** a pre-filled `ARCHITECTURE.md` template
-**So that** I have the right structure without starting from scratch
-
-**Acceptance Criteria:**
-
-- [x] All required sections present (Overview, Layers, Data Model, Decisions, Best Practices, Migration)
-- [x] Layers & dependencies pre-filled (same rules as guide)
-- [x] Key Decisions format: What/Why/Trade-off/Alternatives
-
----
-
-### Prior Stories (Documentation Guidance)
-
-From `001-guides-review-user-stories.md` → `### architecture-guide.md`:
-
-| #   | Story                                 | Summary                                        | Status |
-| --- | ------------------------------------- | ---------------------------------------------- | ------ |
-| 1   | Single Comprehensive Architecture Doc | One `ARCHITECTURE.md` per project at root      | ✅     |
-| 2   | Design Doc Simplicity                 | Design docs reference arch doc, no duplication | ✅     |
-| 3   | Decision Matrix                       | Tech/data → arch doc; feature → design doc     | ✅     |
-| 4   | Code References                       | Key patterns linked to files with line ranges  | ✅     |
-| 6   | TDD Workflow Integration              | Architecture doc reviewed before coding        | ✅     |
-| 8   | Triggers to Update Architecture Doc   | New data models, tech choices trigger updates  | ✅     |
-| 9   | Common Mistakes                       | No ADR sprawl; keep one comprehensive doc      | ✅     |
-| 11  | Data Architecture Guidance            | Linked guide for data-heavy projects           | ✅     |
-
----
-
-## Scope
-
-**In scope:**
-
-- Phase 1: Add layers/boundaries section to `architecture-guide.md` + create `templates/architecture-template.md`
-- Phase 2: Add `eslint-plugin-boundaries` to `setup-linting.sh` + remove `--biome` mode (breaking change)
-- Phase 3: Create `prompts/` folder + `arch-review.sh` script + git pre-commit hook
-- Phase 4: CI workflow template `templates/ci/architecture-check.yml`
-
-**Out of scope:**
-
-- Phase 5 (Drift detection) - Deferred
-- Python support - Future phase
-- Biome support (dropped - ESLint only)
-
-## Acceptance Criteria
-
-- [x] `architecture-guide.md` has Layers & Boundaries section with:
-  - [x] Layer definitions table
-  - [x] Allowed dependencies matrix
-  - [x] Edge cases (non-3-layer, brownfield, shared utilities)
-- [x] `templates/architecture-template.md` created with layers/boundaries section
-- [x] `setup-linting.sh` adds `eslint-plugin-boundaries` to all ESLint modes
-- [x] `setup-linting.sh` removes `--biome` mode (breaking change)
-- [x] `architecture-guide.md` includes boundary rule setup instructions
-- [x] `framework/prompts/` folder created (copied to `.safeword/prompts/` by setup)
-- [x] `prompts/quality-review.md` extracted from `setup-quality.sh`
-- [x] `prompts/arch-review.md` - semantic architecture review prompt
-- [x] `scripts/arch-review.sh` - calls Haiku API for arch review
-- [x] Git pre-commit hook runs `eslint` + `arch-review.sh` on staged files
-- [x] `setup-safeword.sh` updated to copy `prompts/` folder + install pre-commit hook
-- [x] `setup-quality.sh` updated to reference `.safeword/prompts/quality-review.md`
-- [x] `templates/ci/architecture-check.yml` GitHub Actions workflow
-- [ ] All tests pass
-- [ ] User confirms completion
-
-## Implementation Notes
-
-**Key Design Decisions:**
-
-- ESLint-only (drop Biome support) - simpler, better IDE integration
-- Boundaries via `eslint-plugin-boundaries` - integrated with existing linting
-- Add to existing `architecture-guide.md` rather than new guide (single source of truth)
-- Prompts return structured JSON (per `llm-prompting.md` best practices)
-- Template allows direct app→infra (note hexagonal alternative)

package/.safeword/tickets/003-reactive-fix-prevention.md

@@ -1,57 +0,0 @@
----
-id: 003
-status: todo
----
-
-# Reactive Fix Prevention
-
-**Goal:** Stop agents from falling into "error-fix-error" loops by forcing architectural awareness and planning before changes.
-
-**Why:** Without guardrails, agents fix reactively—creating cascading problems, ignoring root causes, and violating architectural boundaries.
-
-## Work Log
-
-**Purpose:** Track what you've tried so you don't repeat dead ends or lose context.
-
-**CRITICAL: Re-read this ticket before each significant action to stay on track.**
-
----
-
-- 2025-11-27T12:00:00Z Started: Created ticket based on AI research into reactive fix patterns
-- 2025-11-27T12:00:00Z Research: Reviewed MASAI, ProSEA, Factored Agents papers + DEV.to article
-- 2025-11-27T12:00:00Z Finding: Documentation alone insufficient—need active enforcement + planning phases
-- 2025-11-27T12:00:00Z Created: Research summary at `.safeword/planning/006-reactive-fix-prevention-research.md`
-- 2025-11-27T12:00:00Z Created: User stories at `.safeword/planning/user-stories/003-reactive-fix-prevention.md`
-
----
-
-## Planning Docs
-
-- `.safeword/planning/user-stories/003-reactive-fix-prevention.md`
-- `.safeword/planning/006-reactive-fix-prevention-research.md`
-
-## Scope
-
-**In scope:**
-
-- Add "Before ANY Error Fix" section to SAFEWORD.md
-- Add "Red Flags (Stop and Think)" section to SAFEWORD.md
-- Update systematic-debugging skill with architecture context loading
-- Add reactive fix prevention guidance to code-philosophy.md
-- Sync all changes to packages/cli/templates/
-
-**Out of scope:**
-
-- New tooling (covered by ticket 002)
-- CI integration (covered by ticket 002)
-- Automated change counting hooks (future enhancement)
-
-## Acceptance Criteria
-
-- [ ] User stories complete and validated
-- [ ] "Before ANY Error Fix" section in SAFEWORD.md
-- [ ] "Red Flags (Stop and Think)" section in SAFEWORD.md
-- [ ] systematic-debugging skill updated with architecture context loading step
-- [ ] code-philosophy.md updated with reactive fix prevention
-- [ ] Changes synced to packages/cli/templates/
-- [ ] User confirms completion

package/.safeword/tickets/004-technical-constraints-in-user-stories.md

@@ -1,39 +0,0 @@
----
-id: 004
-status: done
-created: 2025-11-27T14:00:00Z
-last_modified: 2025-11-27T14:30:00Z
----
-
-# Technical Constraints in User Stories
-
-**Goal:** Capture non-functional requirements (NFRs) in user stories before test definitions, so tests are informed by performance, security, and compatibility constraints.
-
-**Why:** Without explicit constraints, agents write tests that verify behavior but miss NFRs—resulting in code that "works" but fails in production (slow, insecure, incompatible).
-
-## Work Log
-
----
-
-- 2025-11-27T14:30:00Z Complete: All files updated, ready for user confirmation
-- 2025-11-27T14:25:00Z Updated: SAFEWORD.md workflow step 1, creation trigger, edge cases
-- 2025-11-27T14:20:00Z Updated: tdd-best-practices.md with example including constraints
-- 2025-11-27T14:15:00Z Updated: user-story-guide.md with constraints guidance section
-- 2025-11-27T14:10:00Z Updated: user-stories-template.md with Technical Constraints section
-- 2025-11-27T14:05:00Z Decision: Option A (add to user stories) vs Option B (separate doc) - chose A for simplicity
-- 2025-11-27T14:00:00Z Started: User asked about capturing technical requirements before test definitions in TDD
-
----
-
-### Planning Docs
-
-- .safeword/planning/user-stories/004-technical-constraints.md
-
-### Acceptance Criteria
-
-- [x] User stories template includes Technical Constraints section
-- [x] User story guide explains how to fill out constraints
-- [x] TDD best practices includes example with constraints
-- [x] SAFEWORD.md workflow mentions constraints
-- [x] Sync framework/SAFEWORD.md to .safeword/SAFEWORD.md
-- [ ] User confirms completion

package/.safeword/tickets/005-cli-implementation.md

@@ -1,248 +0,0 @@
----
-id: 005
-status: pending
-created: 2025-11-27
-github: https://github.com/TheMostlyGreat/safeword/issues/1
----
-
-# Implement TypeScript CLI
-
-**Goal:** Ship TypeScript CLI (`safeword`) that replaces bash scripts with elite developer experience.
-
-**Why:** Bash scripts are hard to maintain, test, and extend. CLI improves DX for setup, verification, and team onboarding.
-
-## Planning Docs
-
-- [CLI UX Vision](../planning/011-cli-ux-vision.md) — All decisions and architecture
-
-## Key Decisions
-
-| Decision             | Choice                                         |
-| -------------------- | ---------------------------------------------- |
-| Package              | `safeword`                                     |
-| Commands             | setup, check, upgrade, diff, reset             |
-| Global flags         | `--version`, `--help`                          |
-| Non-interactive      | Auto-TTY + `--yes`                             |
-| Existing config      | Error: "Run `safeword upgrade`"                |
-| Same-version upgrade | Force reinstall                                |
-| Linting failure      | Core failure, exit 1                           |
-| No git               | Prompt (auto-skip with `--yes`)                |
-| AGENTS.md            | Prepend link (LLM primacy)                     |
-| AGENTS.md check      | SessionStart hook verifies, re-adds if missing |
-| package.json         | Add lint/format scripts                        |
-| Reset confirm        | Prompt (auto-confirm with `--yes`)             |
-| Skill conflict       | Overwrite silently                             |
-| Diff output          | Summary + `--verbose`                          |
-
-## Scope
-
-### v1 Commands
-
-| Command              | Purpose                            |
-| -------------------- | ---------------------------------- |
-| `safeword`           | Show help                          |
-| `safeword --version` | Show CLI version                   |
-| `safeword setup`     | Full setup (error if exists)       |
-| `safeword check`     | Health + versions                  |
-| `safeword upgrade`   | Update project (always reinstalls) |
-| `safeword diff`      | Preview changes                    |
-| `safeword reset`     | Remove (prompts confirm)           |
-
-### Flags
-
-- `--version` — Show CLI version (global)
-- `--help` — Show help (global)
-- `--yes` — Accept defaults (skip prompts, auto-confirm)
-- `--verbose` — Detailed output (diff)
-- `--offline` — Skip version check (check)
-
-### Setup Flow
-
-1. Check existing → error if found
-2. Detect project type
-3. Copy templates to `.safeword/`
-4. Register Claude Code hooks (incl. SessionStart for AGENTS.md check)
-5. Copy skills to `.claude/skills/safeword-*/`
-6. Install + configure linting (exit 1 if fails)
-7. Add lint/format scripts to package.json
-8. Handle git (prompt or auto-skip)
-9. Prepend link to `AGENTS.md` (create if missing)
-
-### Commands on Unconfigured
-
-| Command   | Behavior                     |
-| --------- | ---------------------------- |
-| `check`   | "Not configured" (exit 0)    |
-| `diff`    | Error (exit 1)               |
-| `reset`   | "Nothing to remove" (exit 0) |
-| `upgrade` | Error (exit 1)               |
-
-### Out of Scope (v1.x)
-
-- `safeword doctor`
-- Claude Code plugin wrapper
-- MCP server management
-- Per-package monorepo support
-- CLAUDE.md handling (only AGENTS.md)
-
-## Acceptance Criteria
-
-- [ ] `npx safeword --version` shows version
-- [ ] `npx safeword setup` works (interactive)
-- [ ] `npx safeword setup --yes` works (non-interactive)
-- [ ] Auto-detects no-TTY and uses defaults
-- [ ] Setup errors if `.safeword/` exists
-- [ ] Prompts to init git if missing
-- [ ] Prepends link to `AGENTS.md`
-- [ ] SessionStart hook checks AGENTS.md, re-adds if missing
-- [ ] Full linting setup (exit 1 if fails)
-- [ ] Adds lint/format scripts to package.json
-- [ ] `npx safeword check` shows health + versions
-- [ ] `check` on unconfigured shows "Not configured"
-- [ ] Version check gracefully handles offline
-- [ ] `npx safeword upgrade` always reinstalls
-- [ ] `npx safeword diff` shows summary
-- [ ] `npx safeword diff --verbose` shows full diff
-- [ ] `diff` on unconfigured errors
-- [ ] `npx safeword reset` prompts for confirm
-- [ ] `npx safeword reset --yes` auto-confirms
-- [ ] `reset` on unconfigured says "Nothing to remove"
-- [ ] Reset removes line from `AGENTS.md`
-- [ ] Preserves existing Claude Code hooks
-- [ ] Preserves existing git pre-commit hooks
-- [ ] Refuses to downgrade project
-- [ ] Exit 0 with warnings for partial success
-- [ ] Exit 1 for core failures
-- [ ] Published to npm as `safeword`
-- [ ] README updated
-- [ ] Tested in 3+ real projects
-
-## Technical Notes
-
-### File Structure
-
-```
-.safeword/
-  SAFEWORD.md, version, README.md
-  hooks/, skills/, guides/, scripts/, prompts/, templates/
-
-.claude/
-  settings.json
-  skills/safeword-*/
-
-.git/hooks/
-  pre-commit (if git)
-
-AGENTS.md (link prepended at top)
-```
-
-### Non-Interactive Behavior
-
-- No TTY → auto-defaults
-- `--yes` → force defaults in terminal
-- Defaults: skip git init, auto-confirm reset
-
-### Exit Codes
-
-- 0 = Success (warnings OK)
-- 1 = Core failure
-
-### Hook Coexistence
-
-- Claude: Append/remove by path pattern
-- Git: Marker-based (`SAFEWORD_ARCH_CHECK_START/END`)
-
-### AGENTS.md Handling
-
-- Setup: Prepend link (check for duplicates first)
-- SessionStart: Verify link exists, re-add with alert if missing
-- Reset: Search exact string, remove
-- Only AGENTS.md (not CLAUDE.md)
-
-### package.json Scripts
-
-```json
-{
-  "scripts": {
-    "lint": "eslint .",
-    "format": "prettier --write ."
-  }
-}
-```
-
-**Note:** Reset leaves these intact (standard CLI behavior). User removes manually if wanted.
-
-## Refactor Plan
-
-Minimal refactoring to improve code quality while keeping all 88 tests green.
-
-### 1. Extract Version Utilities
-
-**Why:** `compareVersions` in `upgrade.ts` and `isNewerVersion` in `check.ts` duplicate semver comparison logic.
-
-**Changes:**
-
-- Create `src/utils/version.ts` with:
-  - `compareVersions(a: string, b: string): -1 | 0 | 1`
-  - `isNewerVersion(current: string, latest: string): boolean`
-- Update `upgrade.ts` to import `compareVersions`
-- Update `check.ts` to import `isNewerVersion`
-- Remove duplicate implementations
-
-**Files affected:**
-
-- `src/utils/version.ts` (new)
-- `src/commands/upgrade.ts` (remove local function)
-- `src/commands/check.ts` (remove local function)
-
-### 2. Split Templates File
-
-**Why:** `src/templates/index.ts` at 180+ lines mixes content templates with configuration objects.
-
-**Changes:**
-
-- Create `src/templates/content.ts` for markdown/script content:
-  - `SAFEWORD_MD`, `AGENTS_MD_LINK`, `PRETTIERRC`
-  - `HOOK_AGENTS_CHECK`, `HOOK_PRE_COMMIT`, `HOOK_POST_TOOL`
-  - `SKILL_QUALITY_REVIEWER`
-- Create `src/templates/config.ts` for configuration:
-  - `getEslintConfig()`
-  - `SETTINGS_HOOKS`
-- Update `src/templates/index.ts` to re-export all
-
-**Files affected:**
-
-- `src/templates/content.ts` (new)
-- `src/templates/config.ts` (new)
-- `src/templates/index.ts` (becomes re-export barrel)
-
-### Verification
-
-After each refactor:
-
-1. `npx tsup` - must build successfully
-2. `npx vitest run` - all 88 tests must pass
-
-### Out of Scope
-
-- Command structure changes (not worth the churn)
-- Type improvements for hook filtering (works fine)
-- Adding new abstractions (avoid over-engineering)
-
-### Refactor Checklist
-
-- [ ] Extract version utilities
-- [ ] Verify tests pass
-- [ ] Split templates file
-- [ ] Verify tests pass
-- [ ] Final review
-
-## Work Log
-
-- 2025-11-27 Created ticket
-- 2025-11-27 Added UX vision decisions
-- 2025-11-27 Final: TTY detection, `--yes` defaults, diff format, skill conflicts
-- 2025-11-27 Final: --version, setup errors if exists, linting=exit 1, upgrade reinstalls
-- 2025-11-27 Final: AGENTS.md prepend, SessionStart hook, package.json scripts, reset confirm, unconfigured behavior
-- 2025-11-28 GREEN phase complete: All 88 tests passing
-- 2025-11-28 Added refactor plan

package/.safeword/tickets/006-flesh-out-skills.md

@@ -1,43 +0,0 @@
----
-id: 006
-status: pending
-created: 2025-11-27
-github: https://github.com/TheMostlyGreat/safeword/issues/2
----
-
-# Flesh Out Framework Skills
-
-**Goal:** Expand `framework/skills/` with high-value skills from examples/superpowers.
-
-**Why:** Only `quality-reviewer` exists in framework. Superpowers has 20+ battle-tested skills ready to port.
-
-## Scope
-
-**Priority skills to port:**
-
-| Skill                            | Value                          | Complexity |
-| -------------------------------- | ------------------------------ | ---------- |
-| `systematic-debugging`           | Prevents fix-by-guessing loops | Medium     |
-| `test-driven-development`        | Core TDD workflow              | Low        |
-| `verification-before-completion` | Prevents incomplete work       | Low        |
-| `root-cause-tracing`             | Debug methodology              | Medium     |
-| `writing-plans`                  | Planning discipline            | Low        |
-| `executing-plans`                | Plan execution                 | Low        |
-| `brainstorming`                  | Creative problem solving       | Low        |
-
-**Out of scope:**
-
-- Agent-specific skills (dispatching-parallel-agents, subagent-driven-development)
-- Git workflow skills (using-git-worktrees, finishing-a-development-branch)
-- Meta skills (writing-skills, sharing-skills)
-
-## Acceptance Criteria
-
-- [ ] 5+ skills ported to framework/skills/
-- [ ] Each skill has SKILL.md with clear triggers and protocol
-- [ ] Skills work standalone (no superpowers dependency)
-- [ ] README updated with skills section
-
-## Work Log
-
-- 2025-11-27 Created ticket

package/.safeword/tickets/007-flesh-out-questioning.md

@@ -1,44 +0,0 @@
----
-id: 007
-status: pending
-created: 2025-11-27
-github: https://github.com/TheMostlyGreat/safeword/issues/3
----
-
-# Flesh Out Questioning System
-
-**Goal:** Improve the `ask_questions` functionality in quality review to be more strategic and useful.
-
-**Why:** Current questioning is binary (on/off). Needs guidance on _what_ questions to ask and _when_.
-
-## Current State
-
-- `.auto-quality-review.config` has `ask_questions=true|false`
-- Quality review prompt says "Ask me any non-obvious questions..."
-- No guidance on question types, timing, or prioritization
-
-## Scope
-
-**Improvements:**
-
-1. **Question taxonomy** - Types of questions (clarification, confirmation, scope, edge cases)
-2. **Question triggers** - When to ask vs assume
-3. **Question format** - How to phrase actionable questions
-4. **Question limits** - Max questions per review (avoid question fatigue)
-
-**Deliverables:**
-
-- Questioning guide in `framework/guides/questioning-guide.md`
-- Update quality-reviewer SKILL.md with questioning protocol
-- Update quality review prompts
-
-## Acceptance Criteria
-
-- [ ] Guide defines question types with examples
-- [ ] Clear triggers for when to ask
-- [ ] Limit guidance (e.g., "max 2-3 questions per review")
-- [ ] Quality reviewer SKILL.md references new guide
-
-## Work Log
-
-- 2025-11-27 Created ticket