chore: update agent parity governance to v0.2.0

Author: hindermath

.github/copilot-instructions.md

@@ -297,7 +297,7 @@ For GitLab repositories, use the authenticated `glab` CLI first for equivalent a
 
 ## Spec-Kit-Preset-Pflege / Spec Kit Preset Maintenance
 
-- Standard-Preset-Set: `security-governance` v0.4.0 prio 10, `architecture-governance` v0.2.0 prio 20, `isaqb-architecture-governance` v0.1.0 prio 30, `a11y-governance` v0.2.0 prio 40, `cross-platform-governance` v0.1.0 prio 50, `agent-parity-governance` v0.1.0 prio 60.
+- Standard-Preset-Set: `security-governance` v0.4.0 prio 10, `architecture-governance` v0.2.0 prio 20, `isaqb-architecture-governance` v0.1.0 prio 30, `a11y-governance` v0.2.0 prio 40, `cross-platform-governance` v0.1.0 prio 50, `agent-parity-governance` v0.2.0 prio 60.
 - `security-governance` v0.4.0 fuehrt `AI-SBOM` weiter als bedingt anwendbare Supply-Chain-Evidenz und ergaenzt sprachspezifische Secure-Coding-Profile fuer Rust, Go, Swift, Java/Kotlin, Python und TypeScript/JavaScript. Reine Entwicklungswerkzeug-Nutzung bleibt `N/A`; KI-Runtime-/Produktkomponenten benoetigen Evidenz nach G7/BSI AI-SBOM-Clustern.
 - Alle sechs Presets sind seit 2026-05-04 im `github/spec-kit` Community-Katalog enthalten und liegen zusätzlich als veröffentlichte Repos unter `https://github.com/hindermath/spec-kit-preset-*`.
 - Neue Level-2-Projekte SOLLEN bei der Spec-Kit-Initialisierung die passende Preset-Teilmenge installieren; C#/.NET-Level-2-Projekte verwenden standardmäßig alle sechs Presets, sofern keine begründete Ausnahme dokumentiert ist.
@@ -310,7 +310,7 @@ For GitLab repositories, use the authenticated `glab` CLI first for equivalent a
 - Bei Änderungen an Preset-Regeln immer prüfen, ob `constitution.md`, `.specify/memory/constitution.md`, `AGENTS.md`, `CLAUDE.md`, `GEMINI.md`, `.github/copilot-instructions.md` und relevante Templates ebenfalls aktualisiert werden müssen.
 - Community-/Katalog-Abstimmung läuft über `github/spec-kit#2362`.
 
-*Standard preset set: `security-governance` v0.4.0 prio 10, `architecture-governance` v0.2.0 prio 20, `isaqb-architecture-governance` v0.1.0 prio 30, `a11y-governance` v0.2.0 prio 40, `cross-platform-governance` v0.1.0 prio 50, and `agent-parity-governance` v0.1.0 prio 60. `security-governance` v0.4.0 keeps conditional `AI-SBOM` evidence and adds language-specific secure-coding profiles for Rust, Go, Swift, Java/Kotlin, Python, and TypeScript/JavaScript: development-tool-only AI usage is `N/A`, while AI runtime/product components require G7/BSI AI-SBOM cluster evidence. All six presets are in the `github/spec-kit` community catalog as of 2026-05-04 and are also published under `https://github.com/hindermath/spec-kit-preset-*`. New Level-2 projects should install the applicable subset; C#/.NET Level-2 projects default to all six unless a justified exception is documented. Commit `.specify/presets/` and generated agent/command/skill updates when presets are project policy, but never commit `.specify/presets/.cache/`. Verify installs with `specify preset list`, `specify preset info`, and where relevant `specify preset resolve`. Preset-rule changes require reviewing constitution, all agent guidance files, and relevant templates. Community/catalog coordination happens in `github/spec-kit#2362`.*
+*Standard preset set: `security-governance` v0.4.0 prio 10, `architecture-governance` v0.2.0 prio 20, `isaqb-architecture-governance` v0.1.0 prio 30, `a11y-governance` v0.2.0 prio 40, `cross-platform-governance` v0.1.0 prio 50, and `agent-parity-governance` v0.2.0 prio 60. `security-governance` v0.4.0 keeps conditional `AI-SBOM` evidence and adds language-specific secure-coding profiles for Rust, Go, Swift, Java/Kotlin, Python, and TypeScript/JavaScript: development-tool-only AI usage is `N/A`, while AI runtime/product components require G7/BSI AI-SBOM cluster evidence. All six presets are in the `github/spec-kit` community catalog as of 2026-05-04 and are also published under `https://github.com/hindermath/spec-kit-preset-*`. New Level-2 projects should install the applicable subset; C#/.NET Level-2 projects default to all six unless a justified exception is documented. Commit `.specify/presets/` and generated agent/command/skill updates when presets are project policy, but never commit `.specify/presets/.cache/`. Verify installs with `specify preset list`, `specify preset info`, and where relevant `specify preset resolve`. Preset-rule changes require reviewing constitution, all agent guidance files, and relevant templates. Community/catalog coordination happens in `github/spec-kit#2362`.*
 
 <!-- SPECKIT START -->
 For additional context about technologies to be used, project structure,

.specify/memory/constitution.md

@@ -816,7 +816,7 @@ workspace family consists of:
 | `isaqb-architecture-governance` | `v0.1.0` | `30` | general iSAQB/arc42 architecture governance |
 | `a11y-governance` | `v0.2.0` | `40` | WCAG 2.2 AA, bilingual DE/EN, CEFR B2, inclusive artefacts |
 | `cross-platform-governance` | `v0.1.0` | `50` | Bash/PowerShell parity, macOS/Linux/Windows script governance |
-| `agent-parity-governance` | `v0.1.0` | `60` | synchronized agent guidance across declared AI-agent files |
+| `agent-parity-governance` | `v0.2.0` | `60` | synchronized agent guidance and agent-neutral Spec-Kit model routing across declared AI-agent files |
 
 `security-governance` v0.4.0 keeps conditional `AI-SBOM` evidence and adds language-specific secure-coding profiles for Rust, Go, Swift, Java/Kotlin, Python, and TypeScript/JavaScript: development-tool-only AI usage is `N/A`, while AI runtime/product components require G7/BSI AI-SBOM cluster evidence.
 

.specify/presets/.registry

@@ -227,9 +227,9 @@
       "installed_at": "2026-05-28T20:10:02.071942+00:00"
     },
     "agent-parity-governance": {
-      "version": "0.1.0",
+      "version": "0.2.0",
       "source": "local",
-      "manifest_hash": "sha256:043c8875f717174f0fdf27f61c7e8fcdc8e658f5fddca9142d10e90ce52c060a",
+      "manifest_hash": "sha256:5e0a6d59cb5291092173e9acee873e808f3ec176bcb81a7a913ac300b420796d",
       "enabled": true,
       "priority": 60,
       "registered_commands": {

.specify/presets/agent-parity-governance/README.md

@@ -1,6 +1,6 @@
 # Agent Parity Governance Preset
 
-Version: `0.1.0`
+Version: `0.2.0`
 Requires: `spec-kit >= 0.8.0` (uses the `wrap` and `append` composition
 strategies introduced in 0.8.x).
 
@@ -9,6 +9,8 @@ Purpose:
 - prevent silent process drift between AI-agent guidance files
 - enforce that shared rules land atomically across the project's
   declared AI-agent guidance surfaces
+- keep Spec-Kit model-routing guidance agent-neutral and out of
+  reproducible feature artifacts
 
 Primary source chapter from `home-baseline` constitution:
 
@@ -20,6 +22,7 @@ What it covers:
 - atomic-change discipline (one change → all surfaces)
 - propagation into project templates and `.specify/memory/constitution.md`
 - explicit documentation of intentional deviations
+- agent-neutral Spec-Kit model routing by work type
 - parity-verification artefact (`agent-parity-checklist-template`)
 
 Preset strategy:
@@ -39,6 +42,8 @@ When to use:
   instructions
 - any project where AI-agent surfaces are part of the contributor
   contract
+- any project that wants model routing guidance without pinning a
+  provider-specific model in specs, plans, or tasks
 
 Common surfaces include `AGENTS.md`, `CLAUDE.md`, `GEMINI.md`,
 `.github/copilot-instructions.md`, `.cursorrules`, `.windsurfrules`,
@@ -52,4 +57,4 @@ When not to use:
 
 Recommended standalone install priority:
 
-- `40`
+- `60`

.specify/presets/agent-parity-governance/preset.yml

@@ -3,8 +3,8 @@ schema_version: "1.0"
 preset:
   id: "agent-parity-governance"
   name: "Agent Parity Governance"
-  version: "0.1.0"
-  description: "Adds shared-guidance parity across a project's declared AI-agent instruction surfaces so agent guidance does not drift."
+  version: "0.2.0"
+  description: "Adds shared-guidance parity and agent-neutral Spec-Kit model-routing guidance across a project's declared AI-agent instruction surfaces so agent guidance does not drift."
   author: "Thorsten Hindermann"
   repository: "https://github.com/hindermath/spec-kit-preset-agent-parity-governance"
   license: "MIT"
@@ -72,4 +72,5 @@ tags:
   - "parity"
   - "agent-md"
   - "agent-guidance"
+  - "model-routing"
   - "multi-agent"

.specify/presets/agent-parity-governance/templates/agent-file-addendum.md

@@ -17,3 +17,24 @@
 - Document every `N/A` decision with rationale; never silently omit.
 - Use `agent-parity-checklist-template` as the standard verification
   artefact for shared-guidance changes.
+
+## Spec-Kit Model Routing
+
+- Treat model choice as operational agent-routing guidance, not as a
+  feature requirement.
+- Do not pin model names in `spec.md`, `plan.md`, `tasks.md`, or
+  individual feature specs. Those artifacts must remain reproducible even
+  when model names change or a different AI agent is used.
+- Each agent maps these recommendations to its currently available
+  models. Do not derive a fixed vendor or model requirement from this
+  preset.
+- For Spec-Kit specification, clarification, planning, task generation,
+  and analysis, prefer the strongest available frontier reasoning/coding
+  model.
+- For complete long-running `speckit.implement` or `/speckit-implement`
+  runs, prefer the strongest available long-running agent model; use a
+  frontier model when maximum judgment quality is more important than
+  runtime stability.
+- For focused review or CI fixes, prefer a coding-optimized model.
+- For trivial cleanup, formatting, or low-risk mechanical edits, a fast
+  small coding model is acceptable.

.specify/presets/agent-parity-governance/templates/agent-parity-checklist-template.md

@@ -32,6 +32,15 @@ confirm whether the shared rule landed.
 | Agent guidance addendum template (if maintained) | | |
 | Other template (name): | | |
 
+## Model Routing
+
+| Check | Done? | Notes |
+|-------|-------|-------|
+| Model-routing guidance, if changed, is present in all maintained agent surfaces | | |
+| Guidance is agent-neutral and does not require a specific provider or model | | |
+| `spec.md`, `plan.md`, `tasks.md`, and feature specs contain no model-name requirements | | |
+| Routing distinguishes planning/analyze, long-running implementation, focused fixes, and trivial mechanical edits | | |
+
 ## Intentional Deviations
 
 | Surface | What deviates | Why | Expected lifetime |

.specify/presets/agent-parity-governance/templates/constitution-addendum.md

@@ -37,6 +37,25 @@ surfaces declared by the project are mandatory for that project.
 - Runtime guidance text MUST name all maintained agent surfaces — never
   list only one or two.
 
+### Spec-Kit model routing
+
+Spec-Kit model routing guidance is part of shared agent guidance. It MUST
+stay agent-neutral and MUST NOT pin provider-specific model names in
+feature artifacts.
+
+- `spec.md`, `plan.md`, `tasks.md`, and individual feature specs MUST NOT
+  encode model names as requirements.
+- Each agent SHOULD map the routing guidance to its currently available
+  models.
+- Specification, clarification, planning, task generation, and analysis
+  SHOULD use the strongest available frontier reasoning/coding model.
+- Complete long-running implementation runs SHOULD use the strongest
+  available long-running agent model, unless maximum judgment quality is
+  more important than runtime stability.
+- Focused review and CI fixes SHOULD use a coding-optimized model.
+- Trivial cleanup, formatting, or low-risk mechanical edits MAY use a
+  fast small coding model.
+
 ### Atomic change discipline
 
 A single pull request or commit that touches shared agent guidance MUST

.specify/presets/agent-parity-governance/templates/plan-addendum.md

@@ -8,5 +8,8 @@
   shared principles change.
 - Plan documentation of any intentional deviation between surfaces (what
   deviates, why, expected lifetime).
+- If model-routing guidance changes, keep it agent-neutral and plan the
+  update across all maintained agent surfaces; never pin model names in
+  reproducible Spec-Kit feature artifacts.
 - Plan a parity check task that validates every declared surface reflects
   the same shared rules at the end of the change.

.specify/presets/agent-parity-governance/templates/spec-addendum.md

@@ -8,5 +8,8 @@
   affected.
 - Record whether any intentional deviation between surfaces is required
   and why.
+- If the feature changes Spec-Kit model-routing guidance, record it as an
+  agent-guidance change. Do not write provider-specific model names into
+  feature requirements.
 - Record any justified `N/A` decisions with rationale. Silent omission
   is not allowed.