docs: remove coordinator and mothership terminology from all documentation

[deanq]

Prerequisite: #208

Summary

• Replace hub-and-spoke "coordinator" architecture descriptions with accurate peer-to-peer topology across all documentation
• Remove "coordinator endpoint", "child endpoint", and "primary endpoint" terminology -- all deployed Flash endpoints are equal peers
• Replace mothership_id with flash_environment_id in StateManagerClient examples
• Remove MothershipsProvisioner references, legacy FLASH_IS_MOTHERSHIP env var sections, and outdated Docker testing section
• Delete obsolete VERIFICATION.md
• Clarify that reconciliation happens at deploy time via CLI (flash deploy), not at endpoint boot

Changed files

Architecture docs: Deployment_Architecture.md, Flash_Deploy_Guide.md, Cross_Endpoint_Routing.md, Load_Balancer_Endpoints.md
CLI docs: flash-deploy.md, flash-build.md
Root: README.md
Deleted: VERIFICATION.md

Context

The previous doc round (PR #208) correctly removed old skeleton patterns (main.py, workers/, FastAPI boilerplate) but incorrectly replaced "mothership" with "coordinator", creating a false impression of a hub-and-spoke architecture with a privileged central node. This PR corrects that in two passes:

1. Coordinator removal (d168633): All diagrams and prose now reflect the flat peer topology where endpoints discover each other via State Manager GraphQL.
2. Mothership cleanup (dd22a72): Remaining mothership references removed -- mothership_id → flash_environment_id, MothershipsProvisioner prose removed, FLASH_IS_MOTHERSHIP env var sections dropped, Docker testing section removed, VERIFICATION.md deleted.

Test plan

• make quality-check passes (format, lint, 1033+ tests, 68.70% coverage)
• grep -ri "coordinator" docs/ src/runpod_flash/cli/docs/ README.md returns zero user-facing hits
• grep -ri "mothership" docs/ src/runpod_flash/cli/docs/ README.md returns zero user-facing hits
• No references to "orchestrator", "hub", or "central" in architectural descriptions
• Diagrams show flat peer topology, not hub-and-spoke

[Copilot]

Pull request overview

This PR corrects terminology introduced in PR #208, replacing the incorrect "coordinator" concept with accurate peer-to-peer architecture documentation. All Flash endpoints are equal peers that discover each other via State Manager GraphQL—there is no privileged central coordinator node.

Changes:

• Removed "coordinator", "mothership", and "hub-and-spoke" terminology from all documentation
• Updated architecture diagrams to show flat peer topology instead of hierarchical coordinator model
• Clarified that reconciliation happens at deploy time via CLI (flash deploy), not at endpoint boot
• Scoped FLASH_IS_MOTHERSHIP as a legacy environment variable with no architectural significance in production

Reviewed changes

Copilot reviewed 46 out of 47 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
tests/unit/test_skeleton.py	Updated tests to reflect new flat-file skeleton structure (gpu_worker.py, cpu_worker.py, lb_worker.py)
tests/unit/resources/test_serverless.py	Added tests for live provisioning health check behavior and template exclusion logic
tests/unit/cli/test_run.py	Extensive new tests for auto-discovery, hot-reload, server generation, and numeric directory handling
tests/unit/cli/commands/build_utils/test_path_utilities.py	New test file for path utility functions (file_to_url_prefix, file_to_resource_name, file_to_module_path)
tests/unit/cli/commands/build_utils/test_manifest_mothership.py	Deleted file - mothership detection tests no longer needed
tests/integration/test_run_auto_provision.py	Deleted file - auto-provision tests moved to unit tests
src/runpod_flash/stubs/registry.py	Added CpuLiveLoadBalancer stub registration
src/runpod_flash/stubs/load_balancer_sls.py	Updated to use LiveServerlessMixin instead of specific LiveLoadBalancer class
src/runpod_flash/core/resources/serverless.py	Added live provisioning health check skip, template exclusion logic, and FLASH_MODULE_PATH injection
src/runpod_flash/core/resources/resource_manager.py	Removed verbose URL logging
src/runpod_flash/core/resources/load_balancer_sls_resource.py	Updated logging for LB endpoint deployment
src/runpod_flash/core/api/runpod.py	Updated GraphQL logging messages
src/runpod_flash/client.py	Added LB route handler passthrough logic for direct execution on LB endpoints
src/runpod_flash/cli/commands/run.py	Major rewrite: auto-discovery of @remote functions, .flash/server.py generation, hot-reload via watchfiles
src/runpod_flash/cli/commands/init.py	Updated output to show new flat-file structure
src/runpod_flash/cli/commands/build_utils/scanner.py	Added path utility functions and is_lb_route_handler field
src/runpod_flash/cli/commands/build_utils/manifest.py	Removed mothership detection, added file_path/local_path_prefix/module_path fields
src/runpod_flash/cli/commands/build_utils/lb_handler_generator.py	Removed reconciliation logic from LB handler lifespan
src/runpod_flash/cli/commands/build.py	Removed main.py requirement from project validation
src/runpod_flash/cli/commands/_run_server_helpers.py	New file: helpers for LB route execution in dev server
src/runpod_flash/cli/utils/skeleton_template/*	Complete restructure: removed workers/, main.py, mothership.py; added gpu_worker.py, cpu_worker.py, lb_worker.py
docs/*.md	Updated all architecture documentation to remove coordinator terminology and reflect peer-to-peer topology
src/runpod_flash/cli/docs/*.md	Updated CLI documentation to reflect auto-discovery and new skeleton structure
README.md	Updated quickstart guide to reflect new flat-file structure and auto-discovery
PRD.md	New file: Product Requirements Document for zero-boilerplate experience
scripts/validate-wheel.sh	Updated to validate new skeleton template files

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Pull request overview

Copilot reviewed 47 out of 48 changed files in this pull request and generated no new comments.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.