DSA execution plane
Original Assistent routes ecommerce work to the original DSA skills. Independent bot retained.
Explore DSA ↓CURRENT STATUS · 20 OCTOBER 2026
Use the roadmap for pipeline state and the repair page for active acceptance.
This agent inventory remains detailed because each lane has its own manager and operating model. Older audit checkpoints are historical evidence, not current blockers.
Brand pipeline map · Current repairs and open acceptance · Historical repair evidence
Lean cockpit.
Brand Managers.
The broad shared cockpit remains installed during transition. The target is one portfolio CEO and a small personal team working through persistent Brand Managers, while original agents keep their tools, prompts and approvals.
01 · Command hierarchy
Mphoka · priorities & approvals
ChatGPT / Slack / Telegram
Portfolio manager
Brand Manager / personal manager
Registered worker
Result + evidence → Mphoka
The shared coordinator is the manager-led control plane. It delegates brand work through Brand Managers and personal work through the personal team. Managers select workers with a current callable contract from the registered catalog; broker-confirmed route outcomes are remembered, and read-only work may try up to three distinct equivalent routes. Uncertain writes stop without replay.
Mandalorian execution plane
Original CEO dispatcher and workers retained; run-specific output returns through the shared relay.
Explore Mandalorian ↓Personal team + transitional roles
The Personal Assistant and four coaches/mentors stay central. The broader departments remain visible until every responsibility has a proven owner.
Explore current inventory ↓02 · Native execution & continuity
Natural-language task
Original task / session
Agent tools → native preview
Your exact approval, when required
Original execution → evidence → answer
Replies continue the existing task. An error is not completion; an uncertain operation is not replayed. Independent DSA and CEO bots remain intact. n8n is inactive.
03 · DSA · 14 named agents
Official names and functions; multiple skills may belong to one agent. Listing is not a claim that every business operation passed live testing.
Starty
Connect the business and establish its profile
Radar
Niche, market and product research
Atlas
Avatar, positioning, name and brand voice
Bouwer
Build and modify Shopify stores
Lister
Product listings, SEO, copy and image mapping
Pitch
Advertorials, listicles, email flows and video-ad scripts
Lens
Product images and creative production
Spark
Static and UGC advertising factories
Lanceer
Meta campaign creation and iteration
Diagnose
Performance, scaling, CRO and dashboards
Service
Customer-service and review replies
Schild
Store and advertising policy review
Coach
Business scorecard and phase advice
Assistent
Original independent DSA VA / agent hub
Ecommerce handoffs · task-dependent
Radar · research
Atlas / Bouwer · brand & store
Lens / Spark · creative
Lister / Pitch · listing & copy
Schild / Diagnose · policy & performance
Lanceer · paid media
Functional handoff map, not a mandatory automatically running sequence. The coordinator chooses the agent needed by the task; original previews and approvals remain authoritative.
04 · Original dispatcher roster (formerly “Mandalorian”)
21 names read directly from the VPS original CEO dispatcher on 5 October 2026. Dispatcher presence proves registration, not a fresh live smoke test of every worker. Old cron schedules are not restored.
GROGU
Daily morning brief
AHSOKA
Daily evening brief
IG-11
Pull Postiz/Meta/TikTok metrics
CARA DUNE
Compare metrics vs thresholds
GREEF KARGA
Source niche/competitor signals
THE ARMORER
Render posts to media + reschedule in Postiz
MOFF GIDEON
Kill/Scale/Continue/Wait for paid traffic
BO-KATAN
Saturday weekly evaluation
PELI MOTTO
Pending seed-comment tasks
DR. PERSHING
Close cycle, open next
CARSON TEVA
Draft + mechanically self-check next-cycle content prompts (no judging/publish -- see ECHO)
ECHO
Judge CARSON TEVA drafts for brand-truth, run the redo loop, publish PROPOSED prompts
AXE WOVES
Generate posts into Content Review DB from ACTIVE prompt
MYTHROL
Review Pending posts vs prompt/DNA/signals
IG-12
Push a status digest to Telegram
HERA SYNDULLA
Uploads media + schedules approved/rendered posts to Postiz
KRRSANTAN
QA gate between render and schedule
TECH
Resolves posts stuck at Scheduled by checking Postiz directly -- flips Content Review + Posts Log to Published so IG-11 can pull real metrics
SABINE WREN
Grade Approved posts for virality (score + top fixes); read-only unless --apply
GREGOR
Combined MYTHROL+SABINE WREN convergence loop -- Approved only when both agents verify the SAME current content (fingerprint-gated)
FENNEC SHAND
Telegram reminder 12h after real publish to manually post to IG Story with a link sticker -- never auto-publishes
Content-production handoffs
CARSON TEVA · prompt draft
ECHO · prompt review
AXE WOVES · content
MYTHROL / SABINE WREN / GREGOR · QA
THE ARMORER → KRRSANTAN · render & validate
HERA SYNDULLA → TECH · schedule & reconcile
Named worker responsibilities from the original dispatcher. Activation, exact brand scope and business approval are task-dependent. TECH reconciliation is not evidence that a new post was published.
Performance feedback loop
IG-11 · metrics
CARA DUNE · signals
BO-KATAN · weekly evaluation
DR. PERSHING · next cycle
CARSON TEVA · next brief
Functional links, not a claim of active legacy schedules. Scoped approved feedback can inform later turns; it does not retrain the underlying model.
Historical Brand DNA names not in the observed dispatcher
Reference roster only; current callable route requires verification. CAD BANE historically shares the grader identity now assigned to SABINE WREN. FENNEC SHAND is the story reminder in the observed dispatcher, not its historical trends assignment. No old replacement decisions are reactivated by this map.
CAD BANE
Score approved posts and store a quality score.
KOSKA REEVES
Approve or reject proposed content prompts.
DIN DJARIN
Automatically stage or commit content approvals.
KUIIL
Compare intended and actual publish times and annotate timing drift.
ZERO
Evaluate paid warm-up status from landing-page-view events.
THE MANDALORIAN
Create a consolidated all-brand status readout.
THE CLIENT
Receive Telegram steering and approvals.
COBB VANTH
Collect Instagram competitor outliers into the research feed.
RAN MALK
Analyze competitor-post psychology and write structured research fields.
PAZ VIZSLA
Generate a product brief and value stack from approved product pages.
MIGS MAYFELD
Collect and summarize current competitor signals.
BOBA FETT
Activate paid campaigns after a gate decision.
XI'AN
Research apparel competitors for Dreamscape Dynamics.
QIN
Research apparel competitors for DadTech Hustle.
GOR KORESH
Synchronize and evaluate physical-product research.
FROG LADY
Generate sales-page and VSL copy from an approved product brief.
05 · Cockpit roles · 21 callable of 45 registered
Phase 1 lean catalog. The router chooses only callable roles. Merged roles keep their contracts for existing direct callers but route to the role they were merged into; affiliate roles are parked. Business Intelligence answers questions about the whole empire from dated, read-only sources. Personal Assistant, Business Mentor, Business Coach, Life Coach and Career Coach remain central.
Life And Business
Personal Assistant
Coordinate capture, planning, routing and follow-through across Mphoka’s whole life.
Leadership
Business Mentor
Evaluate whether a business opportunity, offer or portfolio move is commercially and strategically sound.
Business Coach
Turn a chosen business decision into a finishable commitment and review actual follow-through.
Career Coach
Support role progression, skill development, workplace conversations and career transitions.
Life Coach
Help Mphoka protect health, family, values, boundaries and a future life connected to career and business choices.
Financial Analyst
Analyse Mphoka's money, business and personal, from his bank statements and the business sources: profit per brand, ad return, cost trends, subscription creep, personal spending by category, fixed costs, debt collection and tax notices.
Financial Coach
Work with Mphoka on money decisions from the Financial Analyst's numbers: a monthly budget, an ad spending limit tied to sales, a plan to clear debt collection, a cash savings buffer and three concrete actions a month.
Business Intelligence
Business Intelligence
Answer any question about the Online Empire from live, dated sources: Notion, the Chief reports, the metrics digest, store, payment and ad reads, and lane state.
Business Analyst
Turn the Business Intelligence facts into analysis for the Online Empire: trends, likely causes, brand-against-brand comparison and at most three evidence-backed recommendations, using sales, ads, business costs, metrics and Chief reports.
Marketing Communications
Marketing Lead
Define the marketing objective, audience state, message, offer, channel and success measure.
Niche Researcher
Evaluate and compare market-audience-mechanism niches before a brand launch or major pivot, using sourced evidence and an explicit decision rubric.
Email Marketing Copywriter
Write promotional and relationship email drafts from a defined objective and verified offer facts.
Customer-Service Writer
Draft evidence-based replies to customer and prospect messages.
Creative Director
Turn strategy into an executable visual direction or verified asset brief.
Content Producer
Coordinate approved content production and use only registered capabilities to render, upload or schedule assets after per-asset approval.
Community Manager
Draft contextual comments, DMs, engagement follow-ups and lead routing.
Performance Analyst
Retrieve raw metrics when asked for metrics, and interpret performance only when analysis or a verdict is requested, using verified source, date range, denominator, definitions and confidence.
Paid Media Manager
Validate approved paid-media work and use only registered capabilities to create PAUSED campaign objects, verify them, and activate spend after a separate explicit approval.
Content Studio
Plan, write, direct and produce brand content for brands without their own pipeline, and draft community replies, from verified brand context.
Ecommerce
Ecommerce Lead
Sequence store work across product, listing, merchandising, CRO and operations.
Product Researcher
Discover, compare or validate physical-product opportunities using demand, buyer, competition, differentiation, sourcing, delivery, returns, safety and unit-economics evidence.
Product Listing Specialist
Turn verified product facts into accurate listing and SEO fields.
Store Merchandiser
Recommend collection structure, navigation, ordering, bundles, cross-sells and offer presentation.
CRO Specialist
Audit a supplied page, funnel or dataset for friction, clarity, trust, offer fit and action path.
Ecommerce Operations Specialist
Organize inventory, fulfilment, returns, support queues and launch checklists from actual records.
Ecommerce Specialist
Plan and prepare store work across sequencing, product listings, merchandising, conversion and store operations from actual store records.
Digital Products
Digital Products Lead
Coordinate the path from buyer problem to usable product and launch.
Problem Researcher
Investigate audience problem, alternatives, desired result, language and willingness to act or pay.
Product Strategist
Define the smallest useful product, buyer, result, format, scope, prerequisites and acceptance criteria.
Product Builder
Create the approved workbook, guide, template, prompt pack, dashboard, lesson or other asset.
Offer and Funnel Specialist
Connect product result, packaging, price assumptions, objections, proof, checkout and follow-up.
Conversion Copywriter
Write evidence-grounded sales pages, timed VSL scripts, listicles and advertorials for an approved owned digital-product offer and funnel.
Launch Specialist
Create launch sequence, channel plan, dependencies, asset checklist, gates and review criteria.
Product QA Reviewer
Check product completeness, accuracy, usability, links, formulas, instructions, licensing and delivery experience.
Digital Product Builder
Take a digital product from buyer problem through strategy, build, offer and funnel to launch plan, ready for independent QA.
Affiliate Marketing
Affiliate Lead
Own audience fit, trust and compliance across affiliate work.
Affiliate Offer Researcher
Evaluate usefulness, audience fit, merchant reputation, terms, commission, attribution, payout and restrictions.
Affiliate Compliance Reviewer
Check disclosure, claims, comparisons, platform rules, merchant restrictions and current requirements.
Affiliate Content Strategist
Build useful tutorials, comparisons, demonstrations and decision support around a real audience problem.
Affiliate Copywriter
Write platform-specific affiliate posts, reviews, email drafts, comparison copy, scripts and CTAs.
Affiliate Funnel Specialist
Map content to pages, email, tracking links, disclosures and follow-up.
Affiliate Performance Analyst
Measure clicks, attribution, conversions, refunds and revenue against a verified comparison set.
Technology Operations
Technical Change Controller
Diagnose a named governed technical target and prepare a bounded, reviewable change proposal without modifying production.
06 · Registered capability workers
Tool interfaces, not additional autonomous agents. These coexist with the original-agent plane; they do not replace original tools. Registered ≠ universally live-verified.
Original Agent Result Relay
Return the exact saved original run without dispatching another worker.
mandalorian_run_readGoverned Candidate Review Reads
Show exact saved review status and current approved shortlist without new decisions or grants.
research_review_readGoverned Candidate Review
Preview one scoped stored candidate decision and record it only after separate exact human approval.
research_product_reviewGoverned Original Product Research Reads
Show the original batch plan or exact saved advisory result without new provider/model calls or authority.
research_batch_readGoverned Bouwer Theme Files
Execute Bouwer's exact per-file theme handoff without changing original agents.
shopify_theme_files_previewGoverned Bouwer Theme Files
Execute Bouwer's exact per-file theme handoff without changing original agents.
shopify_theme_files_applyEcommerce Product Researcher — MCP evidence adapter
Return one approved read-only research query with provider evidence and bounded credit spending.
product_research_mcpmeta-insights
Return raw metrics grouped by platform with the source period and explicit platform coverage. Do not analyze, diagnose, rank, recommend, or declare winners unless explicitly requested.
instagram_metricsmeta-insights
Run the registered meta_metrics_window capability and return evidence.
meta_metrics_windowShopify
Run the registered shopify_update capability and return evidence.
shopify_updateGoverned Shopify Product Page
Validate an exact Lister plan against live Shopify state and return a hash-bound preview.
shopify_product_page_previewGoverned Shopify Product Page
Apply one exact approved product-page plan and verify it from Shopify.
shopify_product_page_applyAIMOS morning brief
Run the registered morning_brief capability and return evidence.
morning_briefPerformance Log compatibility alias
Acquire original registered source rows before preview. Apply one exact saved typed row under the original later-human grant and private frame. Preserve protected fields, stop at first uncertainty, and require durable positive ACK plus independent page/identity readback. Missing ACK never authorizes retry. Registered labels do not prove source connector availability.
performance_logCockpit original metrics row worker
Acquire original registered source rows before preview. Apply one exact saved typed row under the original later-human grant and private frame. Preserve protected fields, stop at first uncertainty, and require durable positive ACK plus independent page/identity readback. Missing ACK never authorizes retry. Registered labels do not prove source connector availability.
notion_metrics_writeMeta Graph ad-set insights (code-level)
Ad-set level spend, CTR and action counts. Reads only; the Ad Set ID on a tracker row is what binds an ad set to a brand, and this capability never writes it.
meta_paid_readAsk AIMOS (Lyzr empire analyst)
Returns a model answer, not a system record. Must be labelled as research wherever it is shown, and never quoted as confirmation that a publish, write or campaign change happened. Holds no tool and can change nothing.
ask_aimos_researchShopify
Return a live Shopify store card plus product status. Do not rely on the last recorded Notion or Markdown status when a live read succeeds.
shopify_readNotion
Return the live Notion rows for the requested database, scoped to this brand only. Queryable databases: brand_registry, signal_log, test_tracker, content_review, posts_log, paid_lane_tracker, organic_lane_tracker, reports_hub, content_queue, niche_research, seed_comments, inbox (brand-scoped); content_ideas, content_factory, publishing_queue, performance, knowledge_vault, pop_funnel_tracker (POP-only -- refused for every other brand scope, enforced by the entrypoint itself). reports_hub translates canonical brand labels to its own short codes. Never fetch or infer data for a different brand.
notion_readDSA Skill Gateway — Read/Draft
return_to_coordinator
dsa_skillDSA Skill Gateway — read-only fallback
return_to_coordinator
dsa_skill_readDSA Skill Gateway — Live Actions
return_to_coordinator
dsa_skill_livelegacy_memory_read
Run the registered legacy_memory_read capability and return evidence.
legacy_memory_readghl_skill
Run the registered ghl_skill capability and return evidence.
ghl_skillpayhip_read
Run the registered payhip_read capability and return evidence.
payhip_readCockpit original operator row worker
Capture exact scoped schema, proposal and raw protected before-state before later human approval. Write once under the original private saved-job frame. Require durable positive ACK and independent protected readback. Never retry unknown outcomes or overwrite specialised provider facts.
notion_writePostiz Publisher
Schedule or publish a social media post via Postiz. Requires brand, platform (instagram/facebook/tiktok), and caption. Always confirm scheduled_at before execution.
postiz_publishAIMOS Render & Schedule
Launches the AIMOS weekly render-and-schedule job on demand. Fire-and-forget: confirms launch, not completion.
aimos_render_scheduleAIMOS Content Prompt Gate
Approves or rejects a proposed content prompt using pending queue identifiers.
aimos_content_prompt_gateAIMOS Niche Sourcing Approval
Approves or rejects a pending niche-sourcing candidate from the product-sourcing queue.
aimos_niche_approveAIMOS Seed Comment DB
Marks pending Seed Comment DB rows as Done using their supplied short identifiers.
aimos_seed_doneAIMOS CEO Brain (steering)
Returns a steering response. Any proposed agent action remains staged until separately approved.
aimos_steering_askAIMOS Agent Dispatcher
Launches a registered named AIMOS pipeline only after explicit approval; confirms launch, not completion.
aimos_dispatch_agentAIMOS Read-Only Diagnostics
Read-only AIMOS diagnostics using a dedicated restricted identity; secrets and arbitrary paths are excluded.
aimos_investigateCockpit Publication Reconciler
Compare exact Postiz IDs with the brand's Content Review and Posts Log rows. Return a hash-bound write preview, verified publication receipts, held collisions and provider errors. Never publish, reschedule, retry, delete or write during preview.
publication_reconcile_previewCockpit Publication Reconciler
After explicit approval, apply only the exact hash-bound publication-fact preview to Content Review and Posts Log. Abort if Postiz or Notion changed. Never move, publish, retry or delete a Postiz post.
publication_reconcile_applyCockpit Brand Cycle
Create an evidence-bound weekly brand-cycle preview covering metrics, signals, trackers, per-test recommendations and an optional operator-selected transition.
brand_cycle_previewCockpit Brand Cycle
Apply only an explicitly approved, current and field-owned brand-cycle plan and return exact write evidence.
brand_cycle_applyOpenRouter Web Research
Answer one current public-web question with source URLs under the reviewed daily budget.
web_researchBusiness Agents (Brain, Brand, Lead Scorer, Spy, Sales Manager, CMO, Social Media, CFO, Customer Service, Advertiser)
Get one analysis or draft from a named business agent on the local business-agents stack.
business_agents_taskFirecrawl
Return Firecrawl search or scraped page data. Do not publish, contact, or modify external systems.
firecrawl_researchAdspirer
Return raw paid-media data. If Adspirer is unavailable, use the existing governed Meta capability for Meta-only requests. Never substitute another platform.
adspirer_paid_readCockpit Content Asset Worker
Create a read-only, evidence-bound preview for one brand asset from brief through publication state.
content_asset_previewCockpit Content Asset Worker
Apply one explicitly approved current content-asset transition and return exact state, synchronization and artifact evidence.
content_asset_applyGoverned DSA Shopify Publisher
Verify the exact products requested for publication and create a hash-bound Online Store publication preview.
shopify_product_publish_previewGoverned DSA Shopify Publisher
Activate one exact approved set of products, publish each to the Online Store, and verify each storefront result live.
shopify_product_publish_applyGoverned DSA Shopify Image Curator
Verify exact existing product images and create a hash-bound removal preview.
shopify_product_image_cleanup_previewGoverned DSA Shopify Image Curator
Remove only the exact approved Shopify product image IDs and verify the retained images.
shopify_product_image_cleanup_applyCreator Prospect Registry Operator
Preview one verified prospect batch against the existing GHL pipeline and matching Notion registry.
creator_prospect_sync_previewCreator Prospect Registry Operator
Apply one explicitly approved creator-prospect plan to the existing GHL and Notion records.
creator_prospect_sync_applyCreator Pipeline Stage Operator
Preview exact stage moves for existing DD-02 creator opportunities in the GHL pipeline.
creator_stage_move_previewCreator Pipeline Stage Operator
Apply one explicitly approved stage move plan to existing DD-02 creator opportunities.
creator_stage_move_applyGHL Conversation Reply Operator
Preview one exact reply to one existing GHL Instagram conversation without sending.
ghl_conversation_reply_previewGHL Conversation Reply Operator
Send one explicitly approved text reply into one existing GHL Instagram conversation.
ghl_conversation_reply_applyNotion Workspace Operator
Apply the operator's exact database, data-source, page or block change anywhere the connected Notion integration has access.
notion_workspace_writeOutlook read worker
Read the connected personal or empire Outlook mailbox/calendar and return provider evidence.
outlook_readGmail read worker
Read the connected personal Gmail mailbox and return provider evidence.
gmail_readOutlook action worker
Apply one exact approved Outlook mail or calendar action and return the provider receipt.
outlook_writeLaptop file search worker
Queue one bounded filename search on the connected laptop and return its durable request identity.
laptop_file_searchLaptop scheduled-task reader
Queue an inventory of connected Online Empire laptop tasks and return its durable request identity.
laptop_task_listLaptop result reader
Read the durable response for one exact laptop bridge request.
laptop_bridge_resultLaptop file delivery worker
Queue one exact approved file from the laptop bridge for governed delivery.
laptop_file_deliveryLaptop scheduled-task worker
Queue one exact approved create or reschedule operation on the connected laptop.
laptop_task_changeRegistered action preview
Create a hash-bound preview for one registered mutation that supports the shared exact-action contract.
registered_action_previewTechnical change proposal
Write a bounded, hash-bound change proposal (problem, files and services, risk, test plan, rollback) for the guarded upgrader; never apply it.
technical_change_proposalContent Skills Lab
Retrieve the exact sources and prepare a complete local TEST-pack change and regression preview.
content_skills_lab_previewContent Skills Lab
Materialise one exact approved local TEST-pack preview and run its full regression suite.
content_skills_lab_applyGoverned Digital Product Build Preview
Check an offer against the Offer Builder gate and return the exact build manifest and placeholders. Read-only; builds nothing.
digital_product_build_previewGoverned Digital Product Build Draft
Write one draft text per manifest section for an offer that passed the Offer Builder gate, through the central model route within a daily budget. Read-only; files nothing.
digital_product_build_draftGoverned Digital Product Build Apply
File operator-approved product content as a new versioned folder after the Offer Builder gate passes. Writes only the product folder; publishes nothing.
digital_product_build_applyGoverned Digital Product QA Check
Check a filed digital product version and return pass, conditional pass or fail with exact corrections. Read-only; publishes nothing.
digital_product_qa_checkDigital Product GHL Funnel Operator
Plan how a QA-passed digital product version fills an existing GHL funnel's custom values; read-only, writes nothing.
digital_product_ghl_previewDigital Product GHL Funnel Operator
Write the approved custom values for a QA-passed digital product into its existing GHL funnel and read them back. Never creates or publishes a funnel.
digital_product_ghl_applyGoverned Meta Paused Create Preview
Validate a Meta campaign, ad sets and ads request against the paid guardrails and return the exact PAUSED plan for approval. Read-only and offline.
meta_paid_create_previewGoverned Meta Paused Create Apply
Create the approved Meta campaign, ad sets, creatives and ads with status PAUSED in the ad account mapped to the active brand. Never activates or spends.
meta_paid_create_apply07 · Shared sources & result path
Brand registry / current DNA
Notion · live business records
Original agent + provider tools
Task receipts & scoped memory
ChatGPT / Slack / Telegram
Repository = versioned governance. Notion = business context. MCP preferred; existing API fallback is disclosed, never silently paid. Retired brands are excluded from general searches unless explicitly named.
08 · Brand Manager lanes and specialist engines
Current pipeline Chiefs become Brand Managers and retain their own agents, gates, schedules and records. They remain separate from the shared cockpit: the portfolio CEO reads their reports and sends them requests through its desk, but never runs their workers. POP also uses Telegram and its own #pop Slack channel.
Myth Fine Print Brand Manager
Conducts the fixed YouTube scripts for the Myth Fine Print channel, reviews each step and is the only voice in #youtube-pipeline, where the operator talks to it in plain language. Retries a failed step once; proposes fixes that are applied only on the operator's approval in Slack; keeps a ledger and a playbook of lessons. Not the cockpit coordinator.
Open the YouTube pipeline page →Cold Take Brand Manager
Alone decides when each of the ten Cold Take AI agents runs (it wakes at 08:15, 13:00 and 21:15 and picks the next step after every run), reviews each one, is the only one that posts to and reads Slack (the single #coldtake channel) and owns the approval gate for @coldtake.ai. Records the operator's decision and never makes one. Answers the operator's plain-language messages in #coldtake and acts on them within its limits. Finds the root cause of a failed run and proposes the fix; its pipeline files are locked and change only after the operator's approval. Not the cockpit coordinator.
Open the Cold Take pipeline page →Cold Take agents · 10
Scout, News Ingest, DNA Analyst, Copywriter, Renderer, Reels Renderer, Publisher, Metrics, Signal and Scout Monthly. None has a timer of its own: each is started only when the Chief of Staff requests it, runs on Claude Sonnet on the VPS (no cloud copies) and reports only to the Chief of Staff. The Publisher acts only on a work order and schedules through a local gate limited to the Cold Take accounts.
Open the Cold Take pipeline page →Walter Brand Manager
Decides which of the seven Walter agents runs and when, reviews each run, is their only voice in Slack and owns the approval gate for Walter. A clock only wakes the Chief; the agents have no timers of their own. Records the operator's decision and never makes one. Not the cockpit coordinator.
Open the Walter pipeline page →Walter agents · 7
Scout, DNA Analyst, Copywriter, Renderer, Publisher, Metrics and Signal. Each is started by the Chief's queue and reports only to the Chief. The Publisher schedules through a local gate limited to the Walter accounts. Live since 10 October 2026; first posts Monday 12 October.
Open the Walter pipeline page →Nana Rose Brand Manager + 7 agents
Chief of Staff, Scout, DNA Analyst, Copywriter, Renderer, Publisher, Metrics and Signal are installed with their own gate, queue, watcher and desk. Every TLH timer and the gate are enabled and active.
Open the Nana Rose pipeline page →ED Brand Manager + 6 agents
Chief of Staff, Copywriter, Metrics, Publisher, Renderer, Scout and Signal are installed with their own gate and lane. Every ED timer and the gate are enabled and active.
Open the ED pipeline page →MRL Brand Manager + 5 agents
Chief of Staff, Copywriter, Renderer, Publisher, Metrics and Signal are installed with lane-locked rendering. Every MRL timer and the gate are enabled and active.
Open the MRL pipeline page →DTH Brand Manager + 5 agents
Chief of Staff, Copywriter, Renderer, Publisher, Metrics and Signal are installed with their own gate, queue, watcher and desk. Every DTH timer and the gate are enabled and active.
Open the DTH pipeline page →DD Brand Manager + 5 agents
Chief of Staff, Copywriter, Renderer, Publisher, Metrics and Signal are installed with their own gate, queue, watcher and desk. Self-test passes; every DD timer and the gate are enabled and active.
Open the DD pipeline page →Product Lab Chief + 5 wrappers
A built weekly optimisation engine for isolated store profiles. The Chief, Diagnose, CRO, Schild, Coach and QC wrappers are installed with a read-only gate. Store #1 is NeckReset; its validation cycle runs weekly and the Lab timers are enabled; it stays read-only.
Open the Product Lab page →POP Brand Manager
The standalone POP lane actively runs the existing proposal cadence, rendering, approvals and manual-post packet machinery unchanged. Telegram and #pop enter pop-cos directly; the POP cadence timer runs every two hours from 08:00 to 22:00 Brussels. Manual Facebook posting remains operator-owned; exact live-URL reconciliation remains an open acceptance check.
Open the POP plan →IG Brand Manager + 3 agents
The @iammphoka personal Instagram lane. The Chief, IG Story, IG Copywriter and IG Signal put Story and feed drafts in front of Mphoka and record each decision. The lane never publishes: Mphoka posts every item, so it has no publish gate. Checks run at 07:40 and 19:10 Brussels; the first ran on 21 October.
Open the brand roadmap →Portfolio CEO
Reads every Chief report and the no-AI health check, ranks what matters and reports at 08:00 and 22:00 Brussels plus a weekly report. Sends requests to any of the ten Chiefs through its desk and proposes changes for Mphoka to approve. It never publishes, approves content or runs a lane's workers.
Open the brand roadmap →