cfsa-antigravity 2.19.2 → 2.19.4

package/template/.claude/skills/workflows/workflow-setup-workspace-cicd.md

@@ -10,10 +10,6 @@ parameters:
 
 ## Overview
 
-Parity source:
-- `.agent/workflows/setup-workspace-cicd.md`
-
-Port of `.agent/workflows/setup-workspace-cicd.md` to Claude Code skill format.
 
 ## Prerequisites
 

package/template/.claude/skills/workflows/workflow-setup-workspace-data.md

@@ -10,10 +10,6 @@ parameters:
 
 ## Overview
 
-Parity source:
-- `.agent/workflows/setup-workspace-data.md`
-
-Port of `.agent/workflows/setup-workspace-data.md` to Claude Code skill format.
 
 ## Prerequisites
 

package/template/.claude/skills/workflows/workflow-setup-workspace-hosting.md

@@ -10,10 +10,6 @@ parameters:
 
 ## Overview
 
-Parity source:
-- `.agent/workflows/setup-workspace-hosting.md`
-
-Port of `.agent/workflows/setup-workspace-hosting.md` to Claude Code skill format.
 
 ## Prerequisites
 

package/template/.claude/skills/workflows/workflow-setup-workspace-scaffold.md

@@ -10,10 +10,6 @@ parameters:
 
 ## Overview
 
-Parity source:
-- `.agent/workflows/setup-workspace-scaffold.md`
-
-Port of `.agent/workflows/setup-workspace-scaffold.md` to Claude Code skill format.
 
 ## Prerequisites
 

package/template/.claude/skills/workflows/workflow-setup-workspace.md

@@ -12,12 +12,6 @@ parameters:
 
 Runs workspace setup in four ordered shards: scaffold → cicd → hosting → data.
 
-Parity source:
-- `.agent/workflows/setup-workspace.md`
-- `.agent/workflows/setup-workspace-scaffold.md`
-- `.agent/workflows/setup-workspace-cicd.md`
-- `.agent/workflows/setup-workspace-hosting.md`
-- `.agent/workflows/setup-workspace-data.md`
 
 ## Prerequisites
 

package/template/.claude/skills/workflows/workflow-sync-kit.md

@@ -10,18 +10,14 @@ parameters:
 
 ## Overview
 
-Parity source:
-- `.agent/workflows/sync-kit.md`
-
-Port of `.agent/workflows/sync-kit.md`.
 
 ## Step-by-Step
 
-1. Read `.agent/kit-sync.md` state and validate baseline.
+1. Read `.claude/kit-sync.md` state and validate baseline.
 2. Fetch upstream and diff changed files since last synced commit.
 3. Classify changes (net-new, overwrite-safe, semantic-merge required, project-only).
 4. Apply semantic merges for placeholder/value-bearing files.
 5. Copy net-new files and run ideation-structure migration checks.
 6. Audit project-only files for integration gaps.
 7. Validate, review diff, and scan for unresolved placeholders.
-8. Update `.agent/kit-sync.md` with new commit/version/timestamp.
+8. Update `.claude/kit-sync.md` with new commit/version/timestamp.

package/template/.claude/skills/workflows/workflow-update-architecture-map.md

@@ -10,10 +10,6 @@ parameters:
 
 ## Overview
 
-Parity source:
-- `.agent/workflows/update-architecture-map.md`
-
-Port of `.agent/workflows/update-architecture-map.md`.
 
 ## Step-by-Step
 

package/template/.claude/skills/workflows/workflow-validate-phase-quality.md

@@ -10,10 +10,6 @@ parameters:
 
 ## Overview
 
-Parity source:
-- `.agent/workflows/validate-phase-quality.md`
-
-Port of `.agent/workflows/validate-phase-quality.md` to Claude Code skill format.
 
 ## Prerequisites
 

package/template/.claude/skills/workflows/workflow-validate-phase-readiness.md

@@ -10,10 +10,6 @@ parameters:
 
 ## Overview
 
-Parity source:
-- `.agent/workflows/validate-phase-readiness.md`
-
-Port of `.agent/workflows/validate-phase-readiness.md` to Claude Code skill format.
 
 ## Prerequisites
 

package/template/.claude/skills/workflows/workflow-validate-phase.md

@@ -12,10 +12,6 @@ parameters:
 
 Runs validation in two shards: quality → readiness.
 
-Parity source:
-- `.agent/workflows/validate-phase.md`
-- `.agent/workflows/validate-phase-quality.md`
-- `.agent/workflows/validate-phase-readiness.md`
 
 ## Prerequisites
 

package/template/.claude/skills/workflows/workflow-verify-infrastructure.md

@@ -10,10 +10,6 @@ parameters:
 
 ## Overview
 
-Parity source:
-- `.agent/workflows/verify-infrastructure.md`
-
-Port of `.agent/workflows/verify-infrastructure.md`.
 
 ## Step-by-Step
 

package/template/.claude/skills/workflows/workflow-write-architecture-spec-deepen.md

@@ -10,10 +10,6 @@ parameters:
 
 ## Overview
 
-Parity source:
-- `.agent/workflows/write-architecture-spec-deepen.md`
-
-Port of `.agent/workflows/write-architecture-spec-deepen.md` to Claude Code skill format.
 
 ## Prerequisites
 
@@ -49,7 +45,7 @@ Ensure event-consumer cross-references are valid and bidirectional.
 ### Step 4 — Update indexes and tracker
 
 1. Mark shard complete in `docs/plans/ia/index.md`.
-2. Update `.agent/progress/spec-pipeline.md` IA column via protocol.
+2. Update `.claude/progress/spec-pipeline.md` IA column via protocol.
 3. Run conditional bootstrap if shard is the architecture-design stack-lock shard.
 
 ### Step 5 — Ambiguity gates and last-shard branch

package/template/.claude/skills/workflows/workflow-write-architecture-spec-design.md

@@ -10,10 +10,6 @@ parameters:
 
 ## Overview
 
-Parity source:
-- `.agent/workflows/write-architecture-spec-design.md`
-
-Port of `.agent/workflows/write-architecture-spec-design.md` to Claude Code skill format.
 
 ## Prerequisites
 
@@ -24,8 +20,8 @@ Port of `.agent/workflows/write-architecture-spec-design.md` to Claude Code skil
 
 ### Step 0 — Pipeline and map guards
 
-1. Read `.agent/progress/spec-pipeline.md` and select target shard.
-2. Verify required stack map values exist in `.agent/instructions/tech-stack.md`:
+1. Read `.claude/progress/spec-pipeline.md` and select target shard.
+2. Verify required stack map values exist in `.claude/instructions/tech-stack.md`:
    - Databases
    - Security
    - Surfaces

package/template/.claude/skills/workflows/workflow-write-architecture-spec.md

@@ -12,22 +12,18 @@ parameters:
 
 Runs IA specification authoring in two shards: design → deepen.
 
-Parity source:
-- `.agent/workflows/write-architecture-spec.md`
-- `.agent/workflows/write-architecture-spec-design.md`
-- `.agent/workflows/write-architecture-spec-deepen.md`
 
 ## Prerequisites
 
 1. Decomposition outputs exist (`docs/plans/ia/` skeletons + indexes)
-2. `.agent/progress/spec-pipeline.md` exists and includes IA column states
+2. `.claude/progress/spec-pipeline.md` exists and includes IA column states
 3. Stack map is resolved for required categories (Databases, Security, Surfaces)
 
 ## Step-by-Step
 
 ### Step 0 — Pipeline-state shard selection
 
-1. Read `.agent/progress/spec-pipeline.md`.
+1. Read `.claude/progress/spec-pipeline.md`.
 2. Select lowest-numbered IA shard with `not-started` status unless override is explicitly requested.
 3. If no IA shard remains: stop and direct next step to `/write-be-spec`.
 

package/template/.claude/skills/workflows/workflow-write-be-spec-classify.md

@@ -10,15 +10,11 @@ parameters:
 
 ## Overview
 
-Parity source:
-- `.agent/workflows/write-be-spec-classify.md`
-
-Port of `.agent/workflows/write-be-spec-classify.md` to Claude Code skill format.
 
 ## Prerequisites
 
 1. Target IA shard is complete
-2. `docs/plans/ia/index.md` and `.agent/progress/spec-pipeline.md` are present
+2. `docs/plans/ia/index.md` and `.claude/progress/spec-pipeline.md` are present
 
 ## Step-by-Step
 

package/template/.claude/skills/workflows/workflow-write-be-spec-write.md

@@ -10,10 +10,6 @@ parameters:
 
 ## Overview
 
-Parity source:
-- `.agent/workflows/write-be-spec-write.md`
-
-Port of `.agent/workflows/write-be-spec-write.md` to Claude Code skill format.
 
 ## Prerequisites
 
@@ -50,7 +46,7 @@ Apply line-count thresholds:
 ### Step 4 — Index and tracker updates
 
 1. Update `docs/plans/be/index.md`.
-2. Update BE completion state in `.agent/progress/spec-pipeline.md`.
+2. Update BE completion state in `.claude/progress/spec-pipeline.md`.
 3. Handle structural-reference no-spec rows when classification indicates 0 specs.
 
 ### Step 5 — Iterative deepening passes

package/template/.claude/skills/workflows/workflow-write-be-spec.md

@@ -12,22 +12,18 @@ parameters:
 
 Runs BE specification authoring in two shards: classify → write.
 
-Parity source:
-- `.agent/workflows/write-be-spec.md`
-- `.agent/workflows/write-be-spec-classify.md`
-- `.agent/workflows/write-be-spec-write.md`
 
 ## Prerequisites
 
 1. IA shard(s) are complete and marked ✅ in `docs/plans/ia/index.md`
-2. `.agent/progress/spec-pipeline.md` exists
+2. `.claude/progress/spec-pipeline.md` exists
 3. BE conventions/index files exist under `docs/plans/be/`
 
 ## Step-by-Step
 
 ### Step 0 — Pipeline-state shard selection
 
-1. Read `.agent/progress/spec-pipeline.md`.
+1. Read `.claude/progress/spec-pipeline.md`.
 2. Find shards where IA is `complete` and BE is `not-started`.
 3. Select lowest-numbered eligible shard unless user explicitly overrides.
 4. If none remain, stop and recommend `/write-fe-spec`.
@@ -49,7 +45,7 @@ Call skill: `write-be-spec-write`
 Expected outputs:
 - Full BE spec file(s) in `docs/plans/be/`
 - Updated `docs/plans/be/index.md`
-- Updated `.agent/progress/spec-pipeline.md` BE status
+- Updated `.claude/progress/spec-pipeline.md` BE status
 - Ambiguity gate results + next allowed command
 
 ## Quality Gate

package/template/.claude/skills/workflows/workflow-write-fe-spec-classify.md

@@ -10,10 +10,6 @@ parameters:
 
 ## Overview
 
-Parity source:
-- `.agent/workflows/write-fe-spec-classify.md`
-
-Port of `.agent/workflows/write-fe-spec-classify.md` to Claude Code skill format.
 
 ## Prerequisites
 

package/template/.claude/skills/workflows/workflow-write-fe-spec-write.md

@@ -10,10 +10,6 @@ parameters:
 
 ## Overview
 
-Parity source:
-- `.agent/workflows/write-fe-spec-write.md`
-
-Port of `.agent/workflows/write-fe-spec-write.md` to Claude Code skill format.
 
 ## Prerequisites
 
@@ -38,7 +34,7 @@ Apply line-count thresholds:
 ### Step 3 — Index and tracker updates
 
 1. Update `docs/plans/fe/index.md` status.
-2. Update FE completion state in `.agent/progress/spec-pipeline.md` via protocol.
+2. Update FE completion state in `.claude/progress/spec-pipeline.md` via protocol.
 
 ### Step 4 — Iterative deepening passes
 

package/template/.claude/skills/workflows/workflow-write-fe-spec.md

@@ -12,15 +12,11 @@ parameters:
 
 Runs FE specification authoring in two shards: classify → write.
 
-Parity source:
-- `.agent/workflows/write-fe-spec.md`
-- `.agent/workflows/write-fe-spec-classify.md`
-- `.agent/workflows/write-fe-spec-write.md`
 
 ## Prerequisites
 
 1. BE spec for target shard is complete
-2. `.agent/progress/spec-pipeline.md` exists
+2. `.claude/progress/spec-pipeline.md` exists
 3. FE conventions/index files exist under `docs/plans/fe/`
 4. Design system and brand guidance are available for visual surfaces
 
@@ -28,7 +24,7 @@ Parity source:
 
 ### Step 0 — Pipeline-state shard selection
 
-1. Read `.agent/progress/spec-pipeline.md`.
+1. Read `.claude/progress/spec-pipeline.md`.
 2. Find shards where BE is `complete` and FE is `not-started`.
 3. Select lowest-numbered eligible shard unless user explicitly overrides.
 4. If none remain, stop and recommend `/plan-phase`.
@@ -51,7 +47,7 @@ Call skill: `write-fe-spec-write`
 Expected outputs:
 - Full FE spec file(s) in `docs/plans/fe/`
 - Updated `docs/plans/fe/index.md`
-- Updated `.agent/progress/spec-pipeline.md` FE status
+- Updated `.claude/progress/spec-pipeline.md` FE status
 - Ambiguity + navigation completeness gate outcomes
 
 ## Quality Gate

package/template/CLAUDE.md

@@ -14,7 +14,7 @@ This file provides guidance to Claude Code (claude.ai/code) when working with co
 
 ### Installation & Maintenance
 ```bash
-npm run build          # Build template/ from .agent/ and .claude/ for publishing
+npm run build          # Build template/ from the maintained .agent/ and standalone .claude/ trees for publishing
 npm run check          # Verify template integrity
 npm run changeset      # Create a changeset for versioning
 npm test              # Run tests (if present)
@@ -50,7 +50,7 @@ The pipeline enforces a **progressive decision lock** — each stage locks decis
 │   ├── workflows/       # Pipeline workflow skills
 │   ├── setup/           # Setup/bootstrap skills
 │   └── utilities/       # Utility skills
-├── skill-library/       # Shared skill-library (symlink to .agent/skill-library)
+├── skill-library/       # Claude-owned skill library
 ├── rules/               # Always-active constraints
 ├── instructions/        # Core directives (workflow, tech-stack, patterns)
 └── memory/              # Claude-side memory/session continuity
@@ -110,7 +110,7 @@ Templates use `{{PLACEHOLDER}}` markers that are filled by `/bootstrap-agents` (
 
 ### Skill Resolution
 
-Claude workflows resolve skills from the **Surface Stack Map** in `.claude/instructions/tech-stack.md` (source parity with `.agent/` is maintained):
+Claude workflows resolve skills from the **Surface Stack Map** in `.claude/instructions/tech-stack.md`:
 
 1. Determine the shard/slice's surface from its directory path
 2. Look up the row for that surface in the Per-Surface table
@@ -139,7 +139,7 @@ The kit enforces **TDD: Red → Green → Refactor** for all implementation:
 
 ### Non-Negotiable Rules
 
-Rules in `.agent/rules/` apply to **every task**:
+Rules in `.claude/rules/` apply to **every task**:
 
 - **security-first**: PII isolation, input validation, secret handling
 - **tdd-contract-first**: Schemas before implementation, tests ARE the spec
@@ -209,27 +209,27 @@ If ideation-index.md doesn't exist, the pipeline hasn't started — tell user to
 **Every confirmed decision must be written to its output file immediately.** Never batch decisions in-memory across a long conversation. If the conversation truncates, all confirmed work must survive on disk.
 
 This is enforced by:
-- **Session Continuity Protocols** (`.agent/skills/session-continuity/protocols/`)
-- **Progress Tracking** (`.agent/progress/spec-pipeline.md`)
-- **Pattern Extraction** (`.agent/skills/session-continuity/protocols/04-pattern-extraction.md`)
+- **Session Continuity Protocols** (`.claude/skills/session-continuity/protocols/`)
+- **Progress Tracking** (`.claude/progress/spec-pipeline.md`)
+- **Pattern Extraction** (`.claude/skills/session-continuity/protocols/04-pattern-extraction.md`)
 
 ## Common Tasks
 
 ### Adding a New Workflow
 
-1. Create `.agent/workflows/your-workflow.md` (under 12,000 characters)
-2. Create matching Claude workflow skill: `.claude/skills/workflows/workflow-your-workflow.md`
-3. Add slash command shim: `.claude/commands/your-workflow.md`
-4. Ensure workflow file includes `Parity source: .agent/workflows/your-workflow.md`
+1. If you're changing Antigravity, edit `.agent/workflows/your-workflow.md`
+2. If you're changing Claude, edit `.claude/skills/workflows/workflow-your-workflow.md`
+3. If the workflow should exist in both runtimes, add the matching counterpart in the other tree
+4. Add the Claude slash command shim when applicable: `.claude/commands/your-workflow.md`
 5. Update `AGENTS.md`, `GEMINI.md`, and `CLAUDE.md` workflow docs if needed
-6. Run `npm run check` (parity guard enforces .agent↔.claude mapping)
+6. Run `npm run check`
 7. Create changeset: `npm run changeset`
 
 ### Adding a New Skill
 
-1. Create in `.agent/skill-library/[category]/[skill-name]/`
+1. Create it in the runtime you are extending (`.agent/skill-library/[category]/[skill-name]/` or `.claude/skill-library/[category]/[skill-name]/`)
 2. Add `SKILL.md` with YAML frontmatter (name, description)
-3. Add to `.agent/skill-library/MANIFEST.md`
+3. Add it to the matching runtime's `skill-library/MANIFEST.md`
 4. Follow stack-specific pattern if needed (e.g., `references/typescript.md`)
 
 ### Updating Templates

package/template/docs/README.md

@@ -178,20 +178,20 @@ cp cfsa-antigravity/AGENTS.md /path/to/your-project/
 |-------|------------|
 | **Antigravity** | Both `AGENTS.md` and `GEMINI.md` are present — bootstrap fills both during `/create-prd`. |
 | **Gemini CLI** | `GEMINI.md` is your agent config. Bootstrap fills it during `/create-prd`. |
-| **Claude Code** | Copy rules/instructions into `.claude/` using Claude's format |
+| **Claude Code** | Use the standalone `.claude/` runtime installed via `cfsa-antigravity init --agent claude` |
 | **Cursor** | Reference from `.cursorrules` or your Cursor config |
 | **Windsurf** | Reference from `.windsurfrules` or equivalent |
 | **Other** | Follow your agent's convention for loading system instructions |
 
 > [!WARNING]
-> **Important Note on `.gitignore`**
+> **Important Note on agent runtime directories**
 >
-> If you are using AI-powered editors like **Cursor** or **Windsurf**, adding the `.agent/` folder to your `.gitignore` may prevent the IDE from indexing the workflows. This results in slash commands (like `/plan`, `/debug`) not appearing in the chat suggestion dropdown.
+> Do not hide the runtime directory your agent/editor needs to index.
 >
-> **Recommended Solution:** To keep the `.agent/` folder local (not tracked by Git) while maintaining AI functionality:
+> - Antigravity-style installs use `.agent/`
+> - Claude Code installs use `.claude/`
 >
-> 1. Ensure `.agent/` is **NOT** in your project's `.gitignore`.
-> 2. Instead, add it to your local exclude file: `.git/info/exclude`
+> **Recommended Solution:** Keep the installed runtime directory out of shared `.gitignore` rules and use `.git/info/exclude` for local-only exclusions.
 
 ### Start
 
@@ -205,15 +205,15 @@ The pipeline tells you what to run next at every step. You never have to guess.
 
 ## The Skill System
 
-**Bundled skills** ship with the kit in `.agent/skills/`. These are universal capabilities every project needs regardless of tech stack — things like TDD workflow, clean code principles, accessibility auditing, brainstorming, and systematic debugging. They're always present, always loaded.
+**Bundled skills** ship with the active runtime's `skills/` directory. These are universal capabilities every project needs regardless of tech stack — things like TDD workflow, clean code principles, accessibility auditing, brainstorming, and systematic debugging. They're always present, always loaded.
 
-**Skill library** lives in `.agent/skill-library/`. These are stack-specific and surface-specific skills that get provisioned by bootstrap as tech decisions are confirmed during `/create-prd`. They're never loaded directly — bootstrap copies the relevant ones into `.agent/skills/` and fills any placeholders with your project's confirmed values.
+**Skill library** lives in the active runtime's `skill-library/`. These are stack-specific and surface-specific skills that get provisioned by bootstrap as tech decisions are confirmed during `/create-prd`. They're never loaded directly — bootstrap copies the relevant ones into the runtime's `skills/` directory and fills any placeholders with your project's confirmed values.
 
 Bootstrap fires once per confirmed decision — it fills only what was just decided and leaves everything else untouched.
 
 ### Bootstrap & Template System
 
-Instruction files in `.agent/instructions/` are **templates**, not static files — they ship with `{{PLACEHOLDER}}` markers that bootstrap fills as tech decisions are confirmed during `/create-prd`.
+Instruction files in the active runtime's `instructions/` directory are **templates**, not static files — they ship with `{{PLACEHOLDER}}` markers that bootstrap fills as tech decisions are confirmed during `/create-prd`.
 
 | File | Filled by | When |
 |------|-----------|------|