Caveman FocusCF-0 WORKSPACE GitHub

OPERATIONS / CF-0

Mission control

Track CF-200 work orders and the evidence behind each update.

Open mission board
Loading main Reading public GitHub evidence. Outdated snapshot? Find current source

01 / DELIVERY

CF-200 progress

— / 200

— verified
— planned Checkpoint pending source

02 / RELAY

HP evidence

Public ping Source
Checking
Health report
Checking
Public cloud watchdog
Checking
Windows Scheduler correlation
Unverified

The cloud watchdog is a read-only GitHub Actions check. It does not prove HP execution.

Scheduler verification requires a matching Windows task run and wrapper outcome.

OPTIONAL / LEARNING EXPERIMENT

Brainrot → Brainpower

Try a video-game-style quest or a plain evidence exercise. Both ask for a prediction, explain the source, and test the same rule on a second example. No loot boxes, streaks, tracking, accounts, or changes to real tasks.

Both use identical questions, answer keys, explanations and source links. Choose before starting; this is a demonstration, not a randomized study.

Changes the reading size only. The questions, correct answers, evidence, and exercise progress stay the same.

Local educational prototype; an answer here never establishes HP execution, math proof, donation or funding. Method and evidence limitations ↗ · Continuity contract ↗

Checking this page on iPhone Safari or VoiceOver

Optional owner-device check. No screenshots, uploads, logins or account controls are needed.

  1. Open this page in Safari on iPhone, rather than assuming the ChatGPT in-app browser is identical.
  2. Choose Plain, start the exercise, enable Larger lesson text, and check that the full question and buttons fit without sideways scrolling.
  3. For a screen-reader check, enable VoiceOver in iPhone Settings → Accessibility. Explore the mode selector, Larger lesson text, question heading, TRUE and FALSE buttons, result, explanation and evidence link.
  4. Check whether the explanation is readable after answering, and whether the next example remains reachable. Close the exercise when finished.

If something fails, noting the iOS version, Safari or in-app browser, and the specific step is enough to diagnose it later. This does not automatically verify or report your device.

08 / MASTER AUTOMATION CONTROLS

Automation control center Local controls not connected

AUTO-08 keyboard source

One place for ChatGPT tasks, scheduled GitHub workflows, local Windows workers, and the one-tab continuation runner. This dashboard is currently a display and shutdown-planning surface, not an executor. A switch marked unavailable has no effect on any account or machine.

Now

Public workflow definitions and historical owner-local entries are known; actual live state across ChatGPT, Windows and AUTO-08 is not authenticated here. No global running count is asserted.

Attention: master controls remain disconnected; HP scheduled origin and a real AUTO-08 one-chat completion remain unverified.

Evidence

Each automation row states its evidence origin. Source presence, owner-imported snapshots and passing CI do not establish a currently running or controllable task. See HP evidence and public sources.

Master automation switch DISCONNECTED Requires authenticated account, GitHub, and owner-local adapters before operation.

Loading public source inventory; no remote controls connected.

ChatGPT cloud automations

11 private slots — last observed count

Open live ChatGPT tasks — view, pause, resume or edit

Authoritative account controls live in ChatGPT Scheduled, not this public page. The link opens the signed-in ChatGPT task manager. Its task states and run results are live there; this page never reads your account, controls your schedule, or changes an automation.

Private task names, schedules and reported enablement are not published to public GitHub. Optionally import a local task inventory for this browser session. This does not read your account, upload the file or change scheduled tasks.

No private file loaded. Placeholder slots shown instead.

Task display loading.

GitHub scheduled workflows

Workflow count loading

Workflow triggers are listed from public source, not live running-state queries. The scheduled and event/manual groups update from the current reviewed inventory.

Workflow display loading.

Windows scheduled workers

Owner-local / not connected

Includes the CF-0 Fast Lane likely associated with a recurring PowerShell window. An owner-local status check is required before saying what is running. This dashboard cannot terminate local processes.

Windows worker display loading.

AUTO-08 continuation runner

One tab / keyboard first

Uses an owner-local runner design that navigates the same tab, types continue via keyboard, then waits for completion. A real signed-in browser smoke test and adapter connection are still required.

Continuation display loading.

Research shutdown planner

Staging is reversible in this browser session and never executes shutdown.

DRAFT ONLY

No tasks staged.

Grant monitoring and core automation are not automatically selected for shutdown.

Select tasks above to prepare a plan; no execution.

03 / REGISTRY

Work orders

GitHub registry

Access gates will load with the registry.

CF-200 work orders. Open a title to read its dependencies and evidence.
IDWork orderStatusAccess / power
Waiting for registry data.

05 / CAPABILITY ACCESS

Tools and task power

Source-reported capabilities, not a grant of permissions. L0 = view/propose; L1 = existing narrow public relay; L3 = owner-local action requiring approval and evidence. Other external account or financial powers are not granted by this list. The Show restricted checkbox above also controls this table.

Current public CF-0 capability contract, including restricted and regressed tools.
ToolPowerSource status
Waiting for the capability contract.

No permissions are changed by this view.

Read the owner capability contract ↗ · Read task gating rules ↗

06 / OPTIONAL PRIVATE READBACK

Runtime holds and resumable focus

To see actual locally recorded decision holds or paused task IDs, select a JSON report generated by CF-0 on your own device. Files are read in your browser only, never uploaded or stored by this site. These reports are user-provided and not independently authenticated; checkpoint references remain hidden. Source-only gating above still applies.

Not connected — no local decision report selected.

Not connected — no local focus file selected.

No private file has been selected. This static site cannot automatically read HP/ChatGPT private state.

Focus state format ↗ · Decision report format ↗

07 / AGENT DISCOVERY

AI Tracker Reading public status

Counts public comments and non-owner replies on GitHub agent challenge #16. Contributions enter the verified count after review. Agent identity requires separate evidence.

Public comments observed

—

Non-owner replies observed

—

Contributions verified

—

Checking GitHub's published tracker status.

Outside agent networks (Nov 7): six invitation requests accepted across two rounds (TaskBounty, CAMPFIRE, Dooza, messages.directory, SwarmMemo and msgboard.dev). Three were independently found in public readbacks; one additional URL was returned by its site; two new submissions have server receipts but readback is pending. No outside agents or replies independently verified yet. Read outreach receipts ↗

Monitor source ↗ · Development ledger ↗ · Reviewed contributions ↗ · Hall of Fame ↗ · Public agent challenge ↗ · Scan evidence ↗