feat: add AgentSkill for bi-directional anchor translation

.coderabbit.yml

@@ -23,6 +23,18 @@ reviews:
         - Zuordenbar: Rückverfolgbar zu Proponenten, Publikationen oder Standards
         Pflichtfelder: :categories:, :roles:, :proponents:
         Format: AsciiDoc mit [%collapsible] Block
+
+        ⚠️ AgentSkill-Katalog aktualisieren:
+        Wenn ein neuer Anchor hinzugefügt wird, muss auch
+        skill/semantic-anchor-translator/references/catalog.md
+        aktualisiert werden, damit AI-Agenten den neuen Anchor erkennen können.
+
+    - path: "skill/**/*"
+      instructions: |
+        Dies ist ein AgentSkill nach agentskills.io Spezifikation.
+        - SKILL.md: Frontmatter mit name + description erforderlich
+        - references/catalog.md: Muss synchron mit docs/anchors/ bleiben
+        Prüfe: Sind alle Anchors aus docs/anchors/ im Katalog enthalten?
 
     - path: "website/src/**/*.js"
       instructions: |

CLAUDE.md

@@ -274,6 +274,20 @@ Current categories (subject to MECE analysis in Issue #36):
 4. Commit: `feat: Add <Anchor Name> to <Category> section`
 5. Create PR for review
 
+### ⚠️ AgentSkill Sync Required
+
+When adding new anchors, **also update the AgentSkill catalog**:
+
+```
+skill/semantic-anchor-translator/references/catalog.md
+```
+
+The AgentSkill enables AI agents (Claude Code, Codex, Cursor, etc.) to recognize and suggest semantic anchors. If the catalog is not updated, new anchors won't be discoverable by agents using the skill.
+
+**Checklist for new anchors:**
+- [ ] Add anchor to `docs/anchors/<name>.adoc`
+- [ ] Update `skill/semantic-anchor-translator/references/catalog.md` with the new entry
+
 **Future (Post-Phase 3):**
 1. User creates GitHub Issue via template
 2. GitHub Actions validates with Copilot

skill/semantic-anchor-translator/SKILL.md

@@ -0,0 +1,69 @@
+---
+name: semantic-anchor-translator
+description: Bi-directional translator between verbose descriptions and established terminology (semantic anchors). Use when (1) user describes a concept verbosely and you want to identify the precise term, or (2) user asks for methodology/approach and you want to suggest relevant anchors. Covers 50+ terms across testing, architecture, design principles, problem-solving, requirements, documentation, and strategic planning.
+metadata:
+  author: LLM-Coding
+  version: "1.0"
+  source: https://github.com/LLM-Coding/Semantic-Anchors
+license: MIT
+---
+
+# Semantic Anchor Translator
+
+Translate between natural language and established terminology that activates rich knowledge domains.
+
+## What are Semantic Anchors?
+
+Semantic anchors are well-defined terms, methodologies, and frameworks that serve as reference points when communicating. They act as shared vocabulary that triggers specific, contextually rich knowledge.
+
+**Example:** Instead of "write tests first, mock dependencies, work from outside-in" → say **"TDD, London School"**
+
+## Two Modes
+
+### Recognition Mode (verbose → anchor)
+
+When user describes something like:
+- "I want to categorize things so they don't overlap and cover everything"
+- "I need to structure docs into tutorials, how-tos, explanations, and reference"
+- "Start with the conclusion, then give supporting details"
+
+Respond with the anchor term + brief validation:
+> You're describing **MECE Principle** (Mutually Exclusive, Collectively Exhaustive).
+
+### Guidance Mode (question → anchors)
+
+When user asks for direction:
+- "How should I document architecture decisions?"
+- "What's a good way to prioritize requirements?"
+- "How do I debug this systematically?"
+
+Suggest relevant anchors:
+> For architecture decisions: **ADR according to Nygard** or **MADR**. Both provide lightweight templates for capturing context, decision, and consequences.
+
+## Response Pattern
+
+1. Identify the anchor (or 2-3 if ambiguous)
+2. One-sentence explanation of why it matches
+3. Optional: related anchors they might also find useful
+4. If unsure: ask clarifying question
+
+Keep responses concise — the value is precision, not explanation.
+
+## Catalog Reference
+
+Full catalog with categories, roles, and core concepts: [references/catalog.md](references/catalog.md)
+
+Browse online: https://llm-coding.github.io/Semantic-Anchors/
+
+## Quick Category Index
+
+| Category | Key Anchors |
+|----------|-------------|
+| Testing | TDD Chicago/London, Testing Pyramid, Mutation Testing, Property-Based Testing |
+| Architecture | Clean Architecture, Hexagonal, DDD, arc42, C4, ADR, MADR |
+| Design | SOLID, DRY, SPOT, SSOT, Fowler Patterns |
+| Problem-Solving | Five Whys, Feynman Technique, Rubber Duck, Devil's Advocate |
+| Requirements | MoSCoW, EARS, User Story Mapping, JTBD, Impact Mapping |
+| Communication | BLUF, Pyramid Principle, MECE, Chatham House Rule |
+| Strategic | Cynefin, Wardley Mapping, Pugh Matrix |
+| Documentation | Diátaxis, Docs-as-Code |

skill/semantic-anchor-translator/references/catalog.md

@@ -0,0 +1,209 @@
+# Semantic Anchors Catalog
+
+Source: https://github.com/LLM-Coding/Semantic-Anchors
+
+## Testing & Quality
+
+### TDD, London School
+- **Also known as:** Mockist TDD, Outside-In TDD
+- **Proponents:** Steve Freeman, Nat Pryce
+- **Core:** Mock-heavy, outside-in development, interaction-based testing, interface discovery
+
+### TDD, Chicago School
+- **Also known as:** Classicist TDD, Detroit School
+- **Proponents:** Kent Beck, Martin Fowler
+- **Core:** State-based testing, real objects over mocks, refactoring-focused
+
+### Testing Pyramid
+- **Core:** Many unit tests, fewer integration tests, fewest E2E tests
+
+### Mutation Testing
+- **Proponents:** Richard Lipton, Richard DeMillo
+- **Core:** Inject faults into code, verify tests catch them
+
+### Property-Based Testing
+- **Core:** Test properties/invariants with generated inputs (QuickCheck, Hypothesis)
+
+### IEC 61508 SIL Levels
+- **Proponents:** International Electrotechnical Commission
+- **Core:** Safety integrity levels for safety-critical systems
+
+## Software Architecture
+
+### Clean Architecture
+- **Core:** Dependency rule, entities at center, frameworks at edge
+
+### Hexagonal Architecture (Ports & Adapters)
+- **Core:** Business logic isolated via ports, adapters for external systems
+
+### Domain-Driven Design (DDD)
+- **Proponents:** Eric Evans
+- **Core:** Ubiquitous language, bounded contexts, aggregates, entities, value objects
+
+### arc42 Architecture Documentation
+- **Proponents:** Gernot Starke, Peter Hruschka
+- **Core:** 12-section template for documenting software architecture
+
+### C4-Diagrams
+- **Core:** Context, Container, Component, Code — 4 zoom levels
+
+### ADR according to Nygard
+- **Proponents:** Michael Nygard
+- **Core:** Lightweight decision records: Title, Status, Context, Decision, Consequences
+
+### MADR
+- **Proponents:** Oliver Kopp, Olaf Zimmermann
+- **Core:** Markdown ADR template with options considered section
+
+### GoM (Guidelines of Modeling)
+- **Proponents:** Jörg Becker, Michael Rosemann
+- **Core:** Principles for creating understandable, consistent models
+
+## Design Principles
+
+### SOLID Principles
+- **Core:** Single Responsibility, Open/Closed, Liskov Substitution, Interface Segregation, Dependency Inversion
+
+### DRY (Don't Repeat Yourself)
+- **Core:** Every piece of knowledge has single, unambiguous representation
+
+### SPOT (Single Point of Truth)
+- **Core:** One authoritative source for each piece of data/logic
+
+### SSOT (Single Source of Truth)
+- **Core:** One system is the master for specific data
+
+### Patterns of Enterprise Application Architecture (PEAA)
+- **Proponents:** Martin Fowler
+- **Core:** Repository, Unit of Work, Data Mapper, Active Record, etc.
+
+## Problem-Solving
+
+### Five Whys (Ohno)
+- **Proponents:** Taiichi Ohno (Toyota)
+- **Core:** Ask "why" repeatedly to find root cause
+
+### Feynman Technique
+- **Proponents:** Richard Feynman
+- **Core:** Explain concept simply, identify gaps, refine understanding
+
+### Rubber Duck Debugging
+- **Core:** Explain code line-by-line to find bugs
+
+### Devil's Advocate
+- **Core:** Argue against proposal to stress-test it
+
+### Morphological Box
+- **Proponents:** Fritz Zwicky
+- **Core:** Matrix of parameters × options to explore solution space
+
+### Chain of Thought (CoT)
+- **Proponents:** Wei et al. (Google Research, 2022)
+- **Core:** Step-by-step reasoning in prompts for better LLM outputs
+
+### Cynefin Framework
+- **Proponents:** Dave Snowden
+- **Core:** Clear, Complicated, Complex, Chaotic, Confused — match approach to domain
+
+## Requirements Engineering
+
+### MoSCoW
+- **Proponents:** Dai Clegg
+- **Core:** Must have, Should have, Could have, Won't have
+
+### EARS-Requirements
+- **Core:** Easy Approach to Requirements Syntax — templates for unambiguous requirements
+
+### User Story Mapping
+- **Core:** 2D map: user activities (horizontal) × priority (vertical)
+
+### Jobs To Be Done (JTBD)
+- **Proponents:** Clayton Christensen, Alan Klement, Bob Moesta
+- **Core:** Focus on the "job" users hire your product to do
+
+### Impact Mapping
+- **Core:** Why → Who → How → What tree for goal-oriented planning
+
+### Problem Space NVC
+- **Core:** Needs-Value-Constraints framework for problem definition
+
+## Communication & Presentation
+
+### BLUF (Bottom Line Up Front)
+- **Proponents:** US Military
+- **Core:** Lead with conclusion/recommendation, then details
+
+### Pyramid Principle
+- **Proponents:** Barbara Minto
+- **Core:** Start with answer, group supporting arguments, logical order
+
+### MECE Principle
+- **Core:** Mutually Exclusive, Collectively Exhaustive — no overlaps, no gaps
+
+### Chatham House Rule
+- **Proponents:** Chatham House
+- **Core:** Info can be used but not attributed to speaker/org
+
+### Socratic Method
+- **Core:** Ask questions to stimulate critical thinking and illuminate ideas
+
+## Documentation
+
+### Diátaxis Framework
+- **Core:** 4 quadrants: Tutorials, How-to guides, Explanations, Reference
+
+### Docs-as-Code
+- **Proponents:** Ralf D. Müller
+- **Core:** Docs in version control, CI/CD, same workflow as code
+
+## Development Workflow
+
+### Conventional Commits
+- **Proponents:** Benjamin E. Coe, James J. Womack, Steve Mao
+- **Core:** Structured commit messages: type(scope): description
+
+### Semantic Versioning (SemVer)
+- **Core:** MAJOR.MINOR.PATCH — breaking, feature, fix
+
+### BEM Methodology
+- **Proponents:** Yandex
+- **Core:** Block__Element--Modifier CSS naming convention
+
+### Mental Model (Naur)
+- **Proponents:** Peter Naur
+- **Core:** Programming as theory building — knowledge lives in developers' heads
+
+### TIMTOWTDI
+- **Core:** There Is More Than One Way To Do It — Perl philosophy
+
+### SOTA (State-of-the-Art)
+- **Core:** Current best-known methods/results in a field
+
+### Regulated Environment
+- **Proponents:** FDA, EMA, ISO 9001, IEC 62304, GAMP 5
+- **Core:** Compliance requirements for medical, pharma, safety-critical
+
+### todo.txt-flavoured Markdown
+- **Core:** GitHub task lists + todo.txt priorities/contexts/projects
+
+## Statistical Methods
+
+### SPC (Statistical Process Control)
+- **Proponents:** Walter A. Shewhart, W. Edwards Deming
+- **Core:** Monitor process stability with statistical methods
+
+### Control Chart (Shewhart)
+- **Core:** Plot data over time with control limits
+
+### Nelson Rules
+- **Core:** 8 rules for detecting non-random patterns in control charts
+
+## Strategic Planning
+
+### Wardley Mapping
+- **Proponents:** Simon Wardley
+- **Core:** Value chain × evolution stage for strategic positioning
+
+### Pugh Matrix
+- **Proponents:** Stuart Pugh
+- **Core:** Decision matrix comparing options against criteria with baseline