chore(templates): move the concern directories out of base (#1504) (#1941)

templates/base/<concern>/ becomes templates/<concern>/ for core,
security, workflow, infra, data and language, per ADR-050. Template and
section IDs are unchanged, and every root resolves the same ordered ID
list and the same section directives before and after.

sync.py lists each concern directory on its own and checks base: entries
against the concern directories; smoke's TPL-08 reads the concern
directories directly. Active references follow; history, the frozen
benchmark arms, examples/ and the harness's pinned vendoring path keep
the old paths.

Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>

Author: braboj

CHANGELOG.md

@@ -21,6 +21,13 @@ alters no template carries no entry.
 
 ## [Unreleased]
 
+### Breaking changes
+
+- The concern directories leave `templates/base/`: `templates/base/<concern>/`
+  is now `templates/<concern>/` for core, security, workflow, infra, data and
+  language. Template and section IDs are unchanged; update any path a
+  reference or hybrid file, or a vendored copy, names.
+
 ### Changed
 
 - The repository moved to the `solid-ai-dev` organisation. Generated files

CLAUDE.md

@@ -20,7 +20,12 @@ project type.
 
 ```
 templates/          # All template source files
-  base/             # Cross-cutting rules (core, security, infra, workflow, language)
+  core/             # Cross-cutting foundation (git, docs, quality, testing, ...)
+  security/         # Security rules
+  workflow/         # Session protocol, issues, gates
+  infra/            # CI/CD, containers, deployment
+  data/             # Data modeling, quality, governance, migration
+  language/         # Language-specific rules
   backend/          # Backend layer — HTTP, API, database, observability
   frontend/         # Frontend layer — UX, accessibility, CSS, SSG
   platform/         # CI and security tool mappings per hosting platform
@@ -147,9 +152,11 @@ Stacks without a variant use a bare name (e.g. `htmx.md`).
 ### 2.4 Inheritance model
 
 ```
-base/ ──┬── frontend/ ──┐
-        ├── backend/  ──┼── stack/
-        └── platform/ ──┘
+concerns ──┬── frontend/ ──┐
+           ├── backend/  ──┼── stack/
+           └── platform/ ──┘
+
+concerns = core/, security/, workflow/, infra/, data/, language/
 ```
 
 - Every stack declares `[DEPENDS ON: ...]` at the top
@@ -180,12 +187,12 @@ base/ ──┬── frontend/ ──┐
 ### 2.6 Adding a new base or layer template
 
 1. Create the file in the correct directory:
-   - `templates/base/core/` — foundation (git, docs, quality, etc.)
-   - `templates/base/security/` — security rules
-   - `templates/base/infra/` — CI/CD, containers, deployment
-   - `templates/base/workflow/` — session protocol, issues, gates
-   - `templates/base/language/` — language-specific rules
-   - `templates/base/data/` — data modeling, quality, governance, migration
+   - `templates/core/` — foundation (git, docs, quality, etc.)
+   - `templates/security/` — security rules
+   - `templates/infra/` — CI/CD, containers, deployment
+   - `templates/workflow/` — session protocol, issues, gates
+   - `templates/language/` — language-specific rules
+   - `templates/data/` — data modeling, quality, governance, migration
    - `templates/backend/` — backend services
    - `templates/frontend/` — frontend/UI projects
    - `templates/platform/` — CI platform mappings
@@ -288,7 +295,7 @@ base/ ──┬── frontend/ ──┐
 
 #### Decision logs
 
-- Apply the ADR threshold in `templates/base/core/docs.md`: a record is
+- Apply the ADR threshold in `templates/core/docs.md`: a record is
   owed when a decision changes something a consuming project can observe
   without reading this repository — the composition model. Prose style,
   tooling, release procedure, naming and file layout do not, however
@@ -530,7 +537,7 @@ summarize — visible sequential execution prevents missed steps.
    checks pass
 6. **CLAUDE.md** — for each new convention/rule, apply the
    doc-placement decision tree in
-   `templates/base/workflow/ai-workflow.md` (Doc placement decision
+   `templates/workflow/ai-workflow.md` (Doc placement decision
    tree section): evaluate code → ADR → README → PLAYBOOK →
    CLAUDE.md → memory in order. CLAUDE.md is the home ONLY if the
    agent MUST apply the rule on every turn.

README.md

@@ -123,8 +123,8 @@ effective; declining a new rule needs no ADR or ticket. See the
 <!-- generated:readme-model-limits -->
 | Stack category | Stacks | Largest chain | Prompt | Min context |
 |----------------|--------|---------------|--------|-------------|
-| abstract | 2 | `stack-python-service` — 557K chars | ~159K tokens | 200K |
-| backend | 8 | `stack-django` — 571K chars | ~163K tokens | 200K |
+| abstract | 2 | `stack-python-service` — 556K chars | ~159K tokens | 200K |
+| backend | 8 | `stack-django` — 570K chars | ~163K tokens | 200K |
 | embedded | 1 | `stack-c-embedded` — 363K chars | ~104K tokens | 128K |
 | hypermedia | 1 | `stack-htmx` — 382K chars | ~109K tokens | 128K |
 | library | 3 | `stack-python-lib` — 486K chars | ~139K tokens | 200K |
@@ -211,17 +211,17 @@ meets templates the stacks table does not list.
 | `templates/backend/microservices.md` | backend | Service boundaries, inter-service comms, saga, contract testing |
 | `templates/backend/monitoring.md` | backend | Key metrics, thresholds, alerts, dashboards, incidents |
 | `templates/backend/webhooks.md` | backend | Inbound webhook intake, edge signature verification, fast-ack, retry amplification |
-| `templates/base/workflow/360.md` | workflow | 360-degree project analysis — four stakeholder perspectives, grading |
-| `templates/base/core/agents.md` | core | Output structure, models (inline/reference/hybrid), formatting rules |
-| `templates/base/workflow/ai-workflow.md` | workflow | AI-assisted development lifecycle, work item hierarchy |
-| `templates/base/workflow/communication.md` | workflow | Communication preferences — concise output, shorthand verbs, scope-asking |
-| `templates/base/workflow/compression.md` | workflow | Re-verifying summaries, tables and rollups derived from verified research |
-| `templates/base/data/data-governance.md` | data | Classification, PII handling, retention, ownership, audit trail |
-| `templates/base/data/data-migration.md` | data | Versioned migrations, zero-downtime, rollback strategies |
-| `templates/base/infra/deployment.md` | infra | Deployment targets (cloud/hybrid/offline), certs, LB, registries, secrets |
-| `templates/base/workflow/release.md` | workflow | Semver, version bump propagation, backward compat, cut-over |
-| `templates/base/core/skills.md` | core | Skill authoring — when to write one, frontmatter, triggering, structure, scripts |
-| `templates/base/workflow/two-pass-review.md` | workflow | Split convention checks into mechanical (auto-fix) and contextual (over-flag, judge) passes over one catalogue |
+| `templates/workflow/360.md` | workflow | 360-degree project analysis — four stakeholder perspectives, grading |
+| `templates/core/agents.md` | core | Output structure, models (inline/reference/hybrid), formatting rules |
+| `templates/workflow/ai-workflow.md` | workflow | AI-assisted development lifecycle, work item hierarchy |
+| `templates/workflow/communication.md` | workflow | Communication preferences — concise output, shorthand verbs, scope-asking |
+| `templates/workflow/compression.md` | workflow | Re-verifying summaries, tables and rollups derived from verified research |
+| `templates/data/data-governance.md` | data | Classification, PII handling, retention, ownership, audit trail |
+| `templates/data/data-migration.md` | data | Versioned migrations, zero-downtime, rollback strategies |
+| `templates/infra/deployment.md` | infra | Deployment targets (cloud/hybrid/offline), certs, LB, registries, secrets |
+| `templates/workflow/release.md` | workflow | Semver, version bump propagation, backward compat, cut-over |
+| `templates/core/skills.md` | core | Skill authoring — when to write one, frontmatter, triggering, structure, scripts |
+| `templates/workflow/two-pass-review.md` | workflow | Split convention checks into mechanical (auto-fix) and contextual (over-flag, judge) passes over one catalogue |
 | `templates/platform/github.md` | platform | CodeQL, GitHub Actions, gitleaks action, push protection |
 | `templates/platform/gitlab.md` | platform | Semgrep OSS, GitLab CI/CD, gitleaks CLI |
 | `templates/platform/linear.md` | platform | Label groups, native priority, sub-issues, code-host sync |
@@ -246,7 +246,7 @@ Attach the template directly, or run `python3 tools/resolve.py <root>
 | Generic / other | `AI_CONTEXT.md` |
 <!-- /generated:readme-agents -->
 
-See `templates/base/core/agents.md` for structure, models, and formatting rules.
+See `templates/core/agents.md` for structure, models, and formatting rules.
 
 ## Links
 

docs/ONBOARDING.md

@@ -10,7 +10,8 @@ files. Instead of writing a `CLAUDE.md` or `AGENTS.md` from scratch for every
 project, you compose templates from three layers:
 
 ```
-templates/base/       → rules that apply to every project
+templates/core/       → rules that apply to every project; security/,
+                        workflow/, infra/, data/ and language/ sit beside it
 templates/backend/    → rules for backend services and APIs
 templates/frontend/   → rules for frontend and UI projects
 templates/stack/      → concrete, technology-specific rules
@@ -74,7 +75,7 @@ what differs from the parent.
 
 Example chain:
 ```
-templates/base/core/git.md + templates/base/core/quality.md + ...
+templates/core/git.md + templates/core/quality.md + ...
     ↓
 templates/stack/python-lib.md
     ↓
@@ -115,7 +116,7 @@ Then run the system end-to-end with an agent:
 
 1. Open Claude Code in any project directory
 2. Attach `templates/INTERVIEW.md` and `templates/stack/python-flask.md`
-3. Ask: "Generate a CLAUDE.md for this project using templates/base/core/agents.md
+3. Ask: "Generate a CLAUDE.md for this project using templates/core/agents.md
    format"
 4. Review the output — this is what users get