Guided Folder

.SYSTEMX — The Guided Folder

The Guided Folder is an eight-minute animated walkthrough of keeping a project's goal, plan, tasks, evidence, and next step connected. A fictional community-garden website gives the explanation a concrete task. An auditorium presentation then opens the four original system infographics, with a gardener, audience questions, a laser pointer, arrows, and readable callouts.

Watch the historical 4K film.

Watch the public 4K film on YouTube · Captioned native X video · Current four system maps · Archived original posters · First-time setup · Copy a setup prompt

The film is native 3840 × 2160, approximately 8:00 long, and includes synthetic English narration, English SDH captions, and 13 YouTube chapters. Select captions and playback quality in YouTube as needed. The wiki no longer embeds YouTube's older thumbnail. The controls illustrated inside the animation are part of the explanation; the actual player provides playback and chapter navigation.

What you will learn

Part	Working responsibility	Read next
Global context and Master Plan	Keep the accepted objective, constraints, milestones, and next useful action available	Planning and memory
Canonical tasks and Agent 0	Assign bounded work, integrate evidence, and record acceptance against the actual criteria	Task workflow, Agent 0
Agent X	Preserve original occurrence time, actors, events, attempts, and evidence	Agent X
Native workers and Agent Y	Use host-provided workers or one selected installed external CLI lane with separate scopes and authority	Subagent processing, Agent Y
Agent Z	Apply the same advisory 100-question policy to supplied answers and evidence	Agent Z, 100 questions
ROOT	Compare retained SYSTEMX, harness, and optional bot TODO observations; use real authorized tools and readback to reconcile them	ROOT and task delta
Agent Alpha and AUTOBOT	Select relevant reference detail from ten fixed functions and reuse unchanged results	AUTOBOT and Agent Alpha
DATA and .SYSTEMXP	Keep owner-managed payloads and independent named project records in the selected scope	DATA, SYSTEMX PROJECTS

The folder holds records. Your app, IDE, CLI, or assistant supplies the model, running workers, tools, credentials, and permissions. Reading a role description does not start a process. Plain documents need no runtime; the optional included command tools require Python 3.9 or later.

Visual chapter index: the four original posters

Each numbered entry links to its original film scene and an archived JPEG in immutable alpha.3 source. These are historical 3840 × 2160 export canvases, with no claim of native generated detail. The current four maps and their text descriptions describe the newer beta source. YouTube controls its own chapter-card previews.

1. Full-system overview

Watch this part at 00:43 · Open the archived alpha.3 4K JPEG

Global context, the Master Plan, canonical tasks, agent responsibilities, ROOT, references, DATA, and project scopes connect around one accepted goal.

2. Agent coordination and CLI tooling

Watch this part at 01:08 · Open the archived alpha.3 4K JPEG

Agent 0 coordinates bounded work; the host executes tools, Agent X records evidence, Agent Z gives advisory review, and acceptance stays explicit.

3. AGENTA and the ten AUTOBOTS

Watch this part at 01:36 · Open the archived alpha.3 4K JPEG

Agent Alpha selects useful Design, Soul, Memory, Rules, Research, Brand, Media, Tokens, Data, or Archive references for the current task.

4. DATA and named project scopes

Watch this part at 02:04 · Open the archived alpha.3 4K JPEG

Owner-managed payloads, deliberate access, and named .SYSTEMXP records keep facts, tasks, memory, and evidence in their selected scope.

All 13 chapters

The whole-second links below land inside their intended scenes. They use the corrected starts in the public YouTube description.

Start	Chapter
00:00	Meet the garden project
00:27	Welcome to the coordination folder
00:43	Four infographics: the first task
01:08	Completion, review and acceptance
01:36	AUTOBOT references selected on demand
02:04	Separate projects and deliberate access
02:32	Files first; tools when selected
03:08	Goals, plans, tasks and evidence
04:01	Agent coordination and CLI lanes
05:00	ROOT and the TODO delta
05:40	Agent Alpha and ten references
06:02	DATA, access and .SYSTEMXP
06:59	Prompts, measured inputs and the next turn

Try one useful task

After first-time setup, adapt this prompt to the tools and permissions available in your project:

Use this project's existing .SYSTEMX and its selected record scope.
Read the relevant context, Master Plan, canonical tasks, and saved next step.
Help me finish [ONE CLEAR OUTCOME].
Agree on evidence that would show this task is complete.
Use Agent 0 to coordinate bounded work with the authorized tools available here.
Use X for meaningful time/event evidence, Y only for a selected installed CLI
task, Z for requested advisory review, and Alpha for relevant reference detail.
Reconcile affected harness or bot task items through actual tools and readback.
Save verified results, open questions, and the exact next action.

A larger project benefits from a researched brief and useful phases, milestones, and independent tasks. Expand only the levels needed for the work. See the one-shot project guide for the researched 20-page app example and acceptance gates; use shared workspaces for multiple authorized chats, bots, or storage clients.

Evidence, access, and measurement

The garden is a fictional educational example. The animation is not a live model session, deployed application, external CLI/PR receipt, or autonomous swarm demonstration. Saved ROOT observations do not continuously synchronize external task lists. Agent X records schedules; external tools execute them. Agent Z scores supplied answers and does not perform its hundred checks or accept work. Agent Alpha selects references, rather than starting ten models.

DATA is owner-managed and is not automatically ingested. Git-ignore defaults do not provide encryption, cloud exclusion, or permission to upload. A Drive URL requires actual authorized access and transport. Record changes need their own ownership, conflict handling, and readback.

The film's token illustration uses constructed input counts. It does not establish a measured bill, execution speed, answer quality, or a universal savings percentage. Maintaining context also costs tokens and time. Compare the complete workflow against an ordinary compact handoff, retaining setup overhead and small tasks. See tokens, cost, and time and evidence and acceptance.

Film and software identities

This is media edition 5.0. It demonstrates implemented 1.11.0-alpha.6 behavior. The separate 1.11.0-alpha.7 source-cleanup patch changes current publication files and documentation; it does not add a model, agent runtime, synchronization service, or new execution authority. Use Versions and changelog for the exact released installer version, pins, evidence, and publication status.

The current template keeps video viewing on YouTube, with published PDFs, images, and reusable operating records in the source tree. Film authoring data, Blender projects, audio, captions, and render/preview outputs belong to the separate production archive. Current source cleanup does not erase historical GitHub releases or remove previously adopted local files. GitHub Actions remains disabled. The software is MIT-licensed alpha; use it at your own risk.

Earlier films and the research paper

Reference	Preserved identity
Original introduction	Media edition 1.0; 1:50 / 1080p; alpha.5 feature reference
Advanced walkthrough	Media edition 2.0; 5:54 / 1080p; alpha.6 coordination and separate local input-count fixture
Start with a Folder	Media edition 3.0; 9:55 / native 4K; beginner-paced garden story
White paper edition 2.0	Technical paper examining alpha.4, published with alpha.5
Four-part system infographic	Original static alpha.3 media exports and accessible text

Created by Lucas (SatoshiUNO). A product of Wayne Tech Lab LLC.

Networks.Chat — portfolio · WayneTechLab.com — business · Source repository