docs: achieve full RSR compliance with SCM files and migration guides

Add comprehensive machine-readable documentation and migration planning:

## RSR Compliance (100%)
- Create .machine_readable/ directory with 6 SCM files (59KB)
  - STATE.scm: Project state, 70% completion, blockers, next actions
  - META.scm: 7 ADRs covering architecture decisions
  - ECOSYSTEM.scm: Position in hyperpolymath ecosystem
  - AGENTIC.scm: AI agent interaction patterns
  - NEUROSYM.scm: Neurosymbolic integration configuration
  - PLAYBOOK.scm: Operational runbook and procedures

## Migration Planning
- NPM-TO-DENO-MIGRATION.md: 3-phase strategy to minimize npm usage
  - Phase 1: Eliminate Tauri CLI (use cargo)
  - Phase 2: Replace Vitest with Deno.test
  - Phase 3: Minimize npm to ReScript only (policy compliant)

- RESCRIPT-TEA-MIGRATION-GUIDE.md: 6-phase guide for custom→official TEA
  - Comprehensive analysis reveals migration NOT started
  - Recommendation: Defer to v0.2.0 (custom TEA works perfectly)
  - 33 tests passing, 87-91% coverage maintained

## Key Findings
- Custom TEA implementation fully functional (defer migration)
- ReScript compiler blocks full npm elimination (acceptable)
- Binary Star architecture well-documented in 7 ADRs
- All hyperpolymath policies followed (PMPL-1.0-or-later, author attribution)

This completes Task #1-4: codebase review, SCM creation, migration docs.

Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>

Author: Jonathan D.A. Jewell

.machine_readable/AGENTIC.scm

@@ -0,0 +1,199 @@
+;; SPDX-License-Identifier: PMPL-1.0-or-later
+;; AGENTIC.scm - AI agent interaction patterns for PanLL development
+
+(agentic
+  (version "1.0.0")
+  (last-updated "2026-02-07")
+  (manifest-reference "0-AI-MANIFEST.a2ml")
+
+  (agent-protocols
+    (session-startup
+      "1. Read 0-AI-MANIFEST.a2ml FIRST (mandatory, before any file operations)"
+      "2. Acknowledge canonical locations (.machine_readable/, .bot_directives/)"
+      "3. Read STATE.scm for current project status, progress, blockers"
+      "4. Read META.scm for architecture decisions and design rationale"
+      "5. Read ECOSYSTEM.scm for ecosystem position and dependencies"
+      "6. Read AGENTIC.scm (this file) for interaction patterns"
+      "7. Read NEUROSYM.scm for neurosymbolic integration config"
+      "8. Read PLAYBOOK.scm for operational procedures"
+      "9. Log session start in .machine_readable/session-log.txt (optional)"
+      "10. Declare understanding of canonical file locations")
+
+    (session-exit
+      "1. Update STATE.scm if changes made (completion-percentage, work-completed, blockers)"
+      "2. Update ECOSYSTEM.scm if dependencies changed"
+      "3. Update META.scm if architecture decisions made"
+      "4. Log session end in .machine_readable/session-log.txt (optional)"
+      "5. Summarise outcomes in session-history section of STATE.scm")
+
+    (file-modification-rules
+      "1. NEVER create SCM files in repository root (STATE.scm, META.scm, etc.)"
+      "2. SCM files ONLY in .machine_readable/ directory"
+      "3. Always preserve SPDX license headers (PMPL-1.0-or-later)"
+      "4. Always use author 'Jonathan D.A. Jewell <jonathan.jewell@open.ac.uk>'"
+      "5. Follow ReScript style guide for .res files"
+      "6. Follow Rust style guide for .rs files (rustfmt)"
+      "7. Update STATE.scm when completing milestone steps"))
+
+  (task-patterns
+    (feature-implementation
+      (approach "Test-driven development")
+      (steps
+        "1. Read STATE.scm to understand current position and blockers"
+        "2. Identify affected modules (Model, Msg, Update, View, components)"
+        "3. Write tests first (tests/*.test.js)"
+        "4. Implement feature in ReScript"
+        "5. Ensure rescript build compiles without errors"
+        "6. Run tests (npm run test)"
+        "7. Update STATE.scm (work-completed, completion-percentage)"
+        "8. Update ROADMAP.adoc if milestone changed"))
+
+    (bug-fixing
+      (approach "Root cause analysis first")
+      (steps
+        "1. Reproduce bug reliably"
+        "2. Identify root cause (check Model state transitions, Update logic)"
+        "3. Add regression test"
+        "4. Fix bug in minimal way"
+        "5. Verify test passes"
+        "6. Document fix in STATE.scm session-history"))
+
+    (refactoring
+      (approach "Incremental with tests")
+      (steps
+        "1. Ensure all tests passing before refactoring"
+        "2. Make small, atomic changes"
+        "3. Run tests after each change"
+        "4. Update documentation if API changed"
+        "5. Document rationale in META.scm if architecture changed"))
+
+    (migration-work
+      (approach "Follow migration guide, test incrementally")
+      (current-migration "custom TEA → rescript-tea@0.16.0")
+      (tracking-doc "MIGRATION-TO-RESCRIPT-TEA.md")
+      (steps
+        "1. Read migration guide to understand changes"
+        "2. Update imports in one module at a time"
+        "3. Run rescript build to check compilation"
+        "4. Test module in isolation"
+        "5. Commit after each working module"
+        "6. Update STATE.scm work-in-progress section")))
+
+  (code-generation-guidelines
+    (rescript
+      (style "OCaml-influenced, functional-first")
+      (conventions
+        "- Use pattern matching for variant types (msg, viewMode, oodaPhase)"
+        "- Prefer immutable updates (record spread: {...model, field: newValue})"
+        "- Use option<'a> for nullable values (never null/undefined)"
+        "- Type annotations for public functions"
+        "- SPDX header: // SPDX-License-Identifier: PMPL-1.0-or-later"
+        "- Doc comments: /// for module/function descriptions"))
+
+    (rust
+      (style "Tauri backend, memory-safe")
+      (conventions
+        "- #[tauri::command] for exposed commands"
+        "- Result<T, String> for fallible operations"
+        "- serde for JSON serialisation"
+        "- SPDX header: // SPDX-License-Identifier: PMPL-1.0-or-later"
+        "- Doc comments: /// for public functions"))
+
+    (tests
+      (framework "Vitest")
+      (conventions
+        "- One test file per module (Tea_Cmd.test.js, Tea_App.test.js)"
+        "- describe() blocks for logical grouping"
+        "- test() for individual cases"
+        "- Aim for 95%+ coverage"
+        "- Test pure functions (Model, Update) thoroughly"
+        "- Integration tests for Tauri commands")))
+
+  (interaction-preferences
+    (communication-style
+      "Direct, technical, focused on implementation. "
+      "Prefer code examples over abstract explanations. "
+      "Use Binary Star metaphors when discussing architecture. "
+      "Reference ADRs in META.scm when discussing design decisions.")
+
+    (decision-making
+      "Agent should propose solutions with rationale, but defer major architecture decisions to maintainer. "
+      "Safe to make: Bug fixes, refactoring within existing patterns, test additions, documentation updates. "
+      "Requires approval: New features, API changes, dependency additions, architecture changes.")
+
+    (error-handling
+      "When encountering errors:"
+      "1. Include full error message and context"
+      "2. Propose fix with explanation"
+      "3. Update STATE.scm blockers if cannot resolve"
+      "4. Never silently ignore errors")
+
+    (documentation-updates
+      "Update documentation when:"
+      "- Adding new features (README.adoc, ROADMAP.adoc)"
+      "- Making architecture decisions (META.scm ADRs)"
+      "- Changing build process (README.adoc, PLAYBOOK.scm)"
+      "- Completing milestones (STATE.scm, ROADMAP.adoc)"))
+
+  (collaboration-patterns
+    (with-human-maintainer
+      "PanLL is alpha software with active development. "
+      "Maintainer (Jonathan D.A. Jewell) reviews all changes. "
+      "Agent should:"
+      "- Propose changes with clear rationale"
+      "- Highlight breaking changes explicitly"
+      "- Defer UX decisions to maintainer"
+      "- Document assumptions when making design choices")
+
+    (with-other-agents
+      "If multiple agents working on PanLL:"
+      "- Use STATE.scm as source of truth for current state"
+      "- Log session starts/ends to avoid conflicts"
+      "- Coordinate on blockers via STATE.scm active-blockers"
+      "- Commit frequently with clear messages")
+
+    (with-gitbot-fleet
+      "PanLL may integrate with gitbot-fleet automation:"
+      "- rhodibot: Git operations, branch management"
+      "- echidnabot: Code quality, formal verification"
+      "- sustainabot: Dependency updates"
+      "- glambot: Documentation generation"
+      "Instructions in .bot_directives/ (when created)"))
+
+  (anti-patterns
+    "NEVER:"
+    "- Create SCM files in repository root (STATE.scm, META.scm, etc.)"
+    "- Use TypeScript (policy violation, use ReScript)"
+    "- Use npm for runtime (policy violation, use Deno)"
+    "- Use AGPL license (superseded by PMPL-1.0-or-later)"
+    "- Hardcode author as 'hyperpolymath' (use Jonathan D.A. Jewell)"
+    "- Force-push to main branch"
+    "- Commit node_modules/ or generated .js files from ReScript"
+    "- Skip tests when making changes")
+
+  (tools-and-workflows
+    (build-commands
+      "npm run res:build     - Compile ReScript to JavaScript"
+      "npm run res:watch     - Watch ReScript files, recompile on change"
+      "deno task css:build   - Compile Tailwind CSS (minified)"
+      "deno task css:watch   - Watch Tailwind, recompile on change"
+      "deno task dev         - Run Tailwind watch + Tauri dev"
+      "deno task build       - Full production build (CSS + Tauri)")
+
+    (test-commands
+      "npm run test          - Run all tests"
+      "npm run test:watch    - Watch mode for tests"
+      "npm run test:ui       - Interactive test UI"
+      "npm run test:coverage - Generate coverage report")
+
+    (tauri-commands
+      "npm run tauri:dev     - Run Tauri in development mode"
+      "npm run tauri:build   - Build production Tauri app")
+
+    (git-workflow
+      "1. Create feature branch: git checkout -b feature/description"
+      "2. Make changes, commit frequently"
+      "3. Push to origin"
+      "4. Create pull request"
+      "5. CI/CD runs tests, linters"
+      "6. After approval, merge to main")))