Project_Velocity/.Agent Context/Sprint 1/Project Velocity Pre-Sourik Purge Report.md

Project Velocity Pre-Sourik Purge Report

Prepared: 2026-04-12
Scope: Current Project_Velocity codebase excluding Sourik for scoring, with a separate compatibility pass against the incoming Sourik tree
Purpose: Identify what should be kept, archived, consolidated, or purged before merging Sourik into the main codebase
Action in this report: No deletion. Assessment only.

---

Scoring Model

Score	Meaning
1	Banish. Strong candidate for removal from repo or immediate quarantine.
2	Very low value. Likely generated, stale, insecure, or misleading.
3	Archive out of main repo. Has historical value but should not live in the main working tree.
4	Review aggressively. Possible duplicate, stale path, or unclear ownership.
5	Transitional. Keep only if a near-term owner confirms it.
6	Useful but not core. Keep if it supports current delivery or migration.
7	Important support artifact.
8	Core implementation path.
9	Highly important operational or product-critical code.
10	Do not purge. Source-of-truth, live runtime, or critical integration surface.

---

Executive Summary

The current codebase has five major purge or merge-hardening problems before Sourik is brought in:

1. Repo bloat is severe

   • models/ alone is about 13237.61 MB
   • db assets/ is about 3689.78 MB
   • app/ is about 2588.26 MB
   • Payload/ is about 523.81 MB
   • comfy_engine/ is about 524.58 MB

2. The repo contains generated and duplicated artifacts

   • app/dist/ is about 442.56 MB
   • app/public/models/ is about 437.58 MB
   • app/assets/House Floor Plans/ is about 958.41 MB
   • the iOS tree has duplicated source files under both:
     • iOS/App, iOS/Core, iOS/Features
     • iOS/velocity/velocity/...

3. There are many zero-byte placeholder files

   • especially under agents/, backend/api/, backend/database/, backend/sentinel/, comfy_engine/scripts/, infrastructure/aws_scale/, and infrastructure/blackbox_local/
   • these create false surface area and false merge confidence

4. The repo contains security-sensitive or machine-local artifacts

   • desineuron-l4-node.pem in repo root is unacceptable long-term
   • historical bootstrap scripts contain stale IPs, stale assumptions, and one-off deployment logic

5. The incoming Sourik tree overlaps the most dangerous places

   • backend API surface
   • frontend API client and dashboard/marketing modules
   • Oracle/Catalyst/Sentinel concepts
   • alternative runtime stacks in Go/Python

The highest-priority pre-merge action is not deleting business logic. It is removing ambiguity.

---

Quantitative Findings

Largest top-level paths

Path	Size
models/	13237.61 MB
db assets/	3689.78 MB
app/	2588.26 MB
comfy_engine/	524.58 MB
Payload/	523.81 MB
.Agent Context/	8.92 MB
backend/	0.40 MB
iOS/	0.33 MB
infrastructure/	0.27 MB

Generated or duplicated asset-heavy paths

Path	Size	Assessment
app/dist/	442.56 MB	Generated build output. Should not live in source control as working source.
app/public/models/	437.58 MB	Runtime-facing duplicates of model assets.
app/assets/House Floor Plans/	958.41 MB	Source asset tree with duplicate downstream copies.
db assets/	3689.78 MB	Valuable business/demo data, but too heavy for a clean app repo.
models/	13237.61 MB	Runtime model binaries. Belongs in artifact storage, not Git.
Payload/	523.81 MB	Packaged payload/archive staging. Likely not source-of-truth.

Zero-byte files found

Representative zero-byte paths include:

• agents/docker-compose.agents.yml
• agents/openclaw_gateway/openclaw.json
• agents/openclaw_gateway/workspace/AGENTS.md
• agents/openclaw_gateway/workspace/HEARTBEAT.md
• agents/skills/meta_ads_manager.py
• agents/skills/social_publisher.py
• agents/skills/whatsapp_connector.ts
• backend/api/routes_crm.py
• backend/api/routes_oracle.py
• backend/api/routes_weaver.py
• backend/database/pinecone_client.py
• backend/database/schemas.py
• backend/database/supabase_client.py
• backend/sentinel/face_tracker.py
• backend/sentinel/sentiment_engine.py
• comfy_engine/scripts/auto_term_sheet.py
• comfy_engine/scripts/queue_manager.py
• comfy_engine/workflows/cinematic_wan22_14b.json
• comfy_engine/workflows/dream_weaver_restyle.json
• infrastructure/aws_scale/node1_agents.tf
• infrastructure/aws_scale/node2_rendering.tf
• infrastructure/aws_scale/tailscale_config.sh
• infrastructure/blackbox_local/docker-compose.local.yml
• infrastructure/blackbox_local/setup_gpu_env.sh

These files are more dangerous than absent files because they imply functionality that does not exist.

---

Duplicate Code and Structural Redundancy

iOS duplicate source trees

The following file sets are byte-identical duplicates:

• iOS/App/ContentView.swift

• iOS/velocity/velocity/App/ContentView.swift

• iOS/App/VelocityApp.swift

• iOS/velocity/velocity/App/VelocityApp.swift

• iOS/Core/State/AppStore.swift

• iOS/velocity/velocity/Core/State/AppStore.swift

• iOS/Core/UI/GlassBlurView.swift

• iOS/velocity/velocity/Core/UI/GlassBlurView.swift

• iOS/Core/UI/VelocityTheme.swift

• iOS/velocity/velocity/Core/UI/VelocityTheme.swift

• iOS/Features/Dashboard/DashboardView.swift

• iOS/velocity/velocity/Features/Dashboard/DashboardView.swift

• iOS/Features/Oracle/OracleView.swift

• iOS/velocity/velocity/Features/Oracle/OracleView.swift

• iOS/Features/Sentinel/SentinelView.swift

• iOS/velocity/velocity/Features/Sentinel/SentinelView.swift

• iOS/Features/Settings/SettingsView.swift

• iOS/velocity/velocity/Features/Settings/SettingsView.swift

Interpretation:

• one tree is acting like a clean source mirror
• the other is the real Xcode project path
• keeping both guarantees confusion and future merge drift

Frontend demo/runtime split

There is a real split between:

• polished app shell modules in app/src/components/modules/*
• newer Oracle-specific implementation under app/src/oracle/*
• demo fallback content in app/src/oracle/lib/oracleDemoData.ts

This is not purge-worthy by itself, but it is a clear merge-risk zone because multiple “truths” exist:

• product shell truth
• demo truth
• intended production truth

---

Scored Purge Table

Path / Scope	Type	Score	Recommendation	Reason
desineuron-l4-node.pem	file	1	Banish from repo immediately	Private key material does not belong in source control. This is a security risk, not just clutter.
3.0.0	file	1	Banish	Zero-value stray file with no semantic role.
app/dist/	dir	1	Banish from repo, generate in CI/local only	Build output. Large, derived, and guaranteed merge noise.
agents/docker-compose.agents.yml	file	1	Banish unless populated immediately	Zero-byte placeholder that implies a deployable agents stack that does not exist.
agents/openclaw_gateway/openclaw.json	file	1	Banish or implement	Zero-byte config placeholder.
agents/openclaw_gateway/workspace/AGENTS.md	file	1	Banish	Zero-byte placeholder.
agents/openclaw_gateway/workspace/HEARTBEAT.md	file	1	Banish	Zero-byte placeholder.
agents/skills/meta_ads_manager.py	file	1	Banish or implement elsewhere	Empty skill file creates false feature surface.
agents/skills/social_publisher.py	file	1	Banish or implement elsewhere	Same issue.
agents/skills/whatsapp_connector.ts	file	1	Banish or implement elsewhere	Same issue.
backend/api/routes_crm.py	file	1	Banish or fill before merge	Empty route file conflicts conceptually with live backend ownership.
backend/api/routes_oracle.py	file	1	Banish or redirect to real Oracle router	Empty file directly conflicts with actual Oracle implementation path in backend/oracle/router_v1.py.
backend/api/routes_weaver.py	file	1	Banish or implement	Empty route file.
backend/database/pinecone_client.py	file	1	Banish	Empty placeholder, no runtime value.
backend/database/schemas.py	file	1	Banish	Empty placeholder; current schema truth lives elsewhere.
backend/database/supabase_client.py	file	1	Banish	Empty placeholder and architecturally confusing because current truth favors PostgreSQL-first.
backend/sentinel/face_tracker.py	file	1	Banish	Empty placeholder.
backend/sentinel/sentiment_engine.py	file	1	Banish	Empty placeholder.
comfy_engine/scripts/auto_term_sheet.py	file	1	Banish	Empty placeholder.
comfy_engine/scripts/queue_manager.py	file	1	Banish	Empty placeholder.
comfy_engine/workflows/cinematic_wan22_14b.json	file	1	Banish or replace with real workflow export	Empty workflow file is actively misleading.
comfy_engine/workflows/dream_weaver_restyle.json	file	1	Banish or replace with real workflow export	Same issue.
infrastructure/aws_scale/node1_agents.tf	file	1	Banish or implement in a real infra module	Empty Terraform file is dead weight.
infrastructure/aws_scale/node2_rendering.tf	file	1	Banish or implement in a real infra module	Same issue.
infrastructure/aws_scale/tailscale_config.sh	file	1	Banish	Empty and also contradicts current non-Tailscale operating direction.
infrastructure/blackbox_local/docker-compose.local.yml	file	1	Banish or define properly	Empty infra stub.
infrastructure/blackbox_local/setup_gpu_env.sh	file	1	Banish or define properly	Empty infra stub.
backend_deploy_20260401.tgz	file	2	Archive outside repo	Historical deploy artifact, not source.
Payload/comfy_engine.zip	file	2	Archive outside repo	Packaged artifact. Valuable for history, not for source control.
remote_bootstrap_20260401.sh	file	3	Archive under runbooks or infra-history	Historical one-off bootstrap with time-bound assumptions.
patch_nemoclaw_service_20260401.sh	file	3	Archive	One-off patch script, not durable runtime code.
user_data_bootstrap.sh	file	4	Keep only if still current; otherwise archive to infra-history	Contains useful S3/bootstrap logic, but also bakes old assumptions.
dw_gateway_v2_min.py	file	4	Review for consolidation into backend or comfy service layer	Functional, but sits as a root-level one-off service, not in an owned module.
monitor_nvme.py	file	4	Archive to ops scripts if still useful	Ad hoc operational helper, not core product code.
monitor_qwen.py	file	4	Archive to ops scripts if still useful	Same issue.
test_scp.txt	file	1	Banish	Stray scratch file.
app/public/models/	dir	3	Review for consolidation	Likely runtime-facing duplicates of source model assets.
app/assets/House Floor Plans/	dir	5	Keep as design/source assets, but deduplicate outputs	Real source value exists, but there are duplicate exports in public/ and dist/.
db assets/	dir	5	Move to data repository or artifact store	Valuable for demo and product data, but too heavy and merge-hostile for application repo.
models/	dir	2	Move to S3/artifact store, keep manifest only in repo	Runtime models should not live in Git. Canonical storage should be S3 or controlled local store.
iOS/App, iOS/Core, iOS/Features	dir tree	4	Consolidate with actual Xcode tree	Duplicates the active app tree and guarantees drift.
iOS/velocity/velocity/...	dir tree	8	Keep as likely real app source-of-truth	This appears to be the actual project-backed path tied to Xcode.
app/src/oracle/lib/oracleDemoData.ts	file	5	Keep short-term, mark explicitly as demo-only	Useful for UI continuity, but dangerous if mistaken for live data path.
app/src/lib/oracleQueryClient.ts	file	6	Review against app/src/oracle/lib/oracleApiClient.ts	Potential overlap between legacy Oracle client and new Oracle path.
app/src/oracle/lib/oracleApiClient.ts	file	8	Keep	Newer Oracle contract path.
backend/main.py	file	10	Keep	Current backend integration root.
backend/routers/sentinel.py	file	10	Keep	Live Sentinel backbone.
backend/routers/cctv.py	file	9	Keep	Active auto-mode/CCTV path.
backend/routers/videos.py	file	9	Keep	Active video catalog runtime path.
backend/routers/vault.py	file	9	Keep	Live vault and trackable-link logic.
backend/services/nemoclaw_client.py	file	10	Keep	Current AI runtime integration truth.
backend/services/auto_mode_matcher.py	file	9	Keep	Core Sentinel automation logic.
backend/oracle/router_v1.py	file	9	Keep	Real Oracle implementation path, even if not fully mounted in the main runtime.
backend/oracle/*	dir tree	8	Keep	Important product direction with tests and contracts.
backend/tests/oracle/*	dir tree	8	Keep	High-value protection for a fragile future merge area.
app/src/components/modules/Sentinel.tsx and sentinel/*	module area	9	Keep	Core UI for a real implemented subsystem.
app/src/components/modules/Catalyst.tsx	file	7	Keep, but compare against Sourik marketing/catalyst ideas	Active surface with likely future overlap.
app/src/components/modules/Oracle.tsx + app/src/app/oracle/page.tsx	module area	9	Keep	Core UI architecture for Oracle.
infrastructure/desineuron_ingress/*	dir tree	10	Keep	Live infrastructure truth and runbooks.
infrastructure/ops_control_plane/*	dir tree	10	Keep	Live operational control surface.
.Agent Context/Bibels/*	docs	8	Keep and consolidate	High-value product/infra truth.

---

Strong Retain Set

These are the areas that should be treated as current core code, not purge targets:

• backend/main.py
• backend/routers/* except empty placeholder files in backend/api/
• backend/services/*
• backend/oracle/*
• backend/db/*
• backend/tests/*
• app/src/* excluding clearly demo-only or overlapping legacy Oracle client paths that need review
• infrastructure/desineuron_ingress/*
• infrastructure/ops_control_plane/*
• iOS/velocity/velocity/* as the likely canonical iOS app path
• .Agent Context/Bibels/*

---

Redundant or Misleading Areas

1. The agents subtree

Current assessment:

• almost entirely placeholder surface
• no runtime truth
• no concrete ownership
• high probability of direct collision with Sourik’s actual agents and MCP work

Recommendation:

• do not merge any existing non-Sourik agents/ placeholders forward as if they are real
• either delete them later or move them to a scratch/archive area

2. Duplicate iOS trees

Current assessment:

• exact duplicate code across two trees
• one likely mirrors source
• the nested Xcode-backed tree is the practical source of truth

Recommendation:

• pick one tree before large merges
• otherwise every iOS merge becomes a double-merge

3. Root-level ops scripts

Current assessment:

• some are useful operational history
• none should remain ambiguous root-level “maybe runtime” assets

Recommendation:

• either move to infrastructure/archive/ or formalize them under owned infra modules

---

Compatibility Assessment for Incoming Sourik Merge

This section is based on the shape of the incoming Sourik tree and how it overlaps with the current codebase.

High-Conflict Zones

Incoming Sourik Area	Current Main Area	Risk	Why
Sourik/velocity/api/*.py	backend/main.py, backend/routers/*, backend/api/*	High	Both define backend API surfaces for Catalyst, Sentinel, leads, websockets, and chat/log flows.
Sourik/velocity/frontend/src/lib/api.ts	app/src/lib/api.ts, app/src/oracle/lib/oracleApiClient.ts	High	Conflicting client-side contract layers are likely.
Sourik/velocity/frontend/src/components/...	app/src/components/modules/*	High	Dashboard, marketing, and visual modules are likely to overlap semantically.
Sourik/velocity/internal/oracle/*	backend/oracle/*	High	Two different Oracle implementations or mental models may coexist.
Sourik/velocity/internal/nemoclaw/*	backend/services/nemoclaw_client.py, prompt files	High	Strong conceptual overlap around reasoning/runtime responsibilities.
Sourik/velocity/services/comfyui_service.py	infrastructure/desineuron_ingress/*, ops control plane, current Comfy deployment path	High	Comfy runtime ownership is already established elsewhere.
Sourik/VelocityApp/*	iOS/velocity/velocity/*	Medium-High	iOS naming and feature overlap likely.

Medium-Conflict Zones

Incoming Sourik Area	Current Main Area	Risk	Why
Sourik/velocity/marketing/*	backend/api/routes_catalyst.py, app/src/components/modules/Catalyst.tsx	Medium	Similar domain, different implementation lineage.
Sourik/velocity/mcp/*	current repo has almost no real MCP implementation outside ops/docs	Medium	Less direct code collision, but conceptual ownership must be assigned.
Sourik/velocity/integrations/*	current backend service boundaries	Medium	Could be useful if treated as adapter candidates instead of merged blindly.

Low-Compatibility / High-Archive Areas in Sourik

These should not be merged into main as-is later:

• .pytest_cache
• compiled .pyc
• .env
• local prototype.db
• coverage*
• test output logs
• bundled .exe
• scratch guides and extracted intermediates
• duplicate keys such as desineuron-l4-node.pem

---

Pre-Merge Filter Recommendations

Before merging Sourik, do this sequence:

1. Purge or quarantine score 1-2 paths from the current main tree

   • especially secrets, zero-byte placeholders, build output, and archived binaries

2. Resolve canonical owners

   • iOS canonical tree
   • Oracle backend owner path
   • Comfy/infra owner path
   • AI runtime owner path

3. Mark demo-only code explicitly

   • especially Oracle demo fallback and any mock client layers

4. Move heavy artifacts out of repo

   • models
   • generated builds
   • payload zips
   • bulky demo assets if they are not required in the core app repo

5. Merge Sourik by subsystem, not by tree

   • backend API
   • frontend marketing/dashboard
   • Oracle logic
   • MCP/integrations
   • iOS

6. Do not allow two runtimes for the same responsibility

   • one Oracle backend
   • one Comfy orchestration path
   • one Nemoclaw integration contract
   • one frontend API client contract per domain

---

Final Recommendation

The current codebase is not yet ready for a blind merge with Sourik.

The biggest blockers are:

• fake surface area from empty files
• duplicated iOS source roots
• heavy runtime assets in Git
• root-level operational drift
• overlapping conceptual ownership around Oracle, Nemoclaw, Catalyst, and API contracts

The right move is:

• purge ambiguity first
• then merge subsystem by subsystem

That will save far more time than trying to reconcile two partially overlapping stacks after the fact.