# Appendix: reproducing the observations (/appendices/claude-and-chatgpt-2/appendix-reproducing-the-observations)

---
type: Document
title: "Appendix: reproducing the observations"
description: "The five commands behind the observations, what each showed on November 9, 2026, its limits, a reset test you can run, and the sources for the whole section."
status: stable
order: 904.3
ksor:
  owner: team:panaversity
  audience: [ public ]
  approval:
    by: process:panaversity
    at: 2026-11-17T13:37:19Z
chapter: "F00A"
part: Front
expert_status: required
concepts: []
last_verified: 2026-11-09
sources:
  - id: code-tab
    title: "Claude Desktop: Code tab docs (Anthropic, verified 2026-11-09)"
    resource: https://code.claude.com/docs/en/desktop?facet1=pdf
  - id: author-inspection
    title: "Panaversity's inspection of a cloud Claude session: a Python test run on November 8, 2026, and a process and environment inspection on November 9, 2026"
    resource: "appendix-reproducing-the-observations.md"
  - id: chatgpt-inspection
    title: "ChatGPT's inspection of its own Python and terminal environment, November 9, 2026, as reported to Panaversity"
    resource: "Reported to Panaversity by ChatGPT, November 9, 2026"
  - id: agents-api
    title: "Agents API architecture (OpenAI, verified 2026-11-09)"
    resource: https://developers.openai.com/api/docs/guides/agents-api/architecture
  - id: work-cloud-security
    title: "ChatGPT Work cloud security (OpenAI, verified 2026-11-09)"
    resource: https://learn.chatgpt.com/docs/enterprise/chatgpt-work-cloud-security
  - id: managed-agents
    title: "Claude Managed Agents overview (Anthropic, verified 2026-11-09)"
    resource: https://platform.claude.com/docs/en/managed-agents/overview
  - id: neon-better-auth
    title: "Managed Better Auth is now an OAuth provider, with MCP support (Neon, November 2, 2026, verified 2026-11-09)"
    resource: https://neon.com/blog/managed-better-auth-oauth-provider-mcp
  - id: better-auth-mcp
    title: "Better Auth MCP plugin (Better Auth, verified 2026-11-09)"
    resource: https://better-auth.com/docs/plugins/mcp
  - id: desktop-scheduled-tasks
    title: "Schedule recurring tasks in Claude Code Desktop (Anthropic, verified 2026-11-09)"
    resource: https://code.claude.com/docs/en/desktop-scheduled-tasks
  - id: cloud-scheduled-tasks
    title: "Automate work with scheduled tasks (Anthropic, verified 2026-11-09)"
    resource: https://code.claude.com/docs/en/routines
  - id: cowork-architecture
    title: "Claude Cowork architecture overview (Anthropic, verified 2026-11-09; on 2026-11-16 this address redirected to the Unified Claude security guide)"
    resource: https://support.claude.com/en/articles/14479288-claude-cowork-architecture-overview
  - id: codex-harness
    title: "Unlocking the Codex harness: how we built the App Server (OpenAI, February 2026, verified 2026-11-09)"
    resource: https://openai.com/index/unlocking-the-codex-harness/
  - id: chatgpt-agent
    title: "Introducing ChatGPT agent (OpenAI, July 2025, verified 2026-11-09)"
    resource: https://openai.com/index/introducing-chatgpt-agent/
  - id: whats-new
    title: "ChatGPT: What's new (OpenAI, verified 2026-11-09)"
    resource: https://learn.chatgpt.com/docs/whats-new
  - id: openai-sandboxing
    title: "ChatGPT sandbox docs (OpenAI, verified 2026-11-09)"
    resource: https://learn.chatgpt.com/docs/sandboxing
  - id: new-stack
    title: "OpenAI is folding Codex into the ChatGPT app (The New Stack, July 2026, verified 2026-11-09)"
    resource: https://thenewstack.io/openai-codex-work-atlas/
  - id: claude-projects
    title: "Let Claude coordinate ongoing work with Projects (Anthropic, verified 2026-11-09)"
    resource: https://code.claude.com/docs/en/claude-projects
  - id: projects-redesigned
    title: "Projects redesigned: from folder to conversation (Anthropic blog, September 17, 2026, verified 2026-11-09)"
    resource: https://claude.com/blog/projects-redesigned
  - id: cloud-environments
    title: "Configure cloud environments (Anthropic, verified 2026-11-09)"
    resource: https://code.claude.com/docs/en/cloud-environments
  - id: self-hosted
    title: "Self-hosted environments (Anthropic, verified 2026-11-09)"
    resource: https://code.claude.com/docs/en/self-hosted-environments
  - id: self-hosted-identity
    title: "Verify session identity in self-hosted environments (Anthropic, verified 2026-11-09)"
    resource: https://code.claude.com/docs/en/self-hosted-environments-identity
  - id: cross-session
    title: "Message your other Claude Code sessions (Anthropic, verified 2026-11-09)"
    resource: https://code.claude.com/docs/en/cross-session-messaging
  - id: claude-tag-identity
    title: "How agent identity works, Claude Tag (Anthropic, verified 2026-11-09)"
    resource: https://claude.com/docs/claude-tag/concepts/agent-identity
  - id: claude-tag-security
    title: "Security and data handling, Claude Tag (Anthropic, verified 2026-11-09)"
    resource: https://claude.com/docs/claude-tag/concepts/security-and-data
  - id: claude-code-cloud
    title: "Use Claude Code in the cloud (Anthropic, verified 2026-11-09)"
    resource: https://code.claude.com/docs/en/claude-code-on-the-web
generated:
  at: 2026-11-17T13:37:19Z
  by: process:panaversity
trust_tier: unverified
build_id: sha256:b1b8ba6179b4acd06d17bd32e6d6ba124becd3c1e125ab480accc40ff9d9b160
dirty: true
ksor_version: 0.0.60
---

### The commands

Run these commands in the Bash tool of a cloud Claude session. Print only the variables listed. Never print values whose names contain TOKEN, KEY or SECRET, because those are credentials. The run date for this section was November 9, 2026, at 02:45 UTC. Version strings and variable names are a snapshot from that date, and they will change.

```
# 1. Init process and VM uptime
ps -o pid,etime,args -p 1
# 2. Process ancestry of the shell
p=$$; while [ "$p" -gt 1 ]; do ps -o pid=,ppid=,args= -p $p | cut -c1-160; p=$(ps -o ppid= -p $p | tr -d ' '); done
# 3. Claude Code binary and version
which claude; readlink -f "$(which claude)"; claude --version
# 4. Selected environment variables
env | grep -E '^(CLAUDE_CODE_(REMOTE|ENTRYPOINT|SESSION_ORIGIN|VERSION|EXECPATH)|ANTHROPIC_BASE_URL)='
# 5. Working transcript location
ls ~/.claude/projects
```

Here is what each command does, in plain words:

1. Shows the first program the computer started (PID 1) and how long it has been running (its uptime).
2. Follows the chain of programs that started the shell, one parent at a time (its process ancestry).
3. Finds the `claude` program, follows any link to the real file, and prints its version.
4. Prints only the selected environment variables.
5. Lists the folder where Claude Code keeps its working transcripts.

### What each step showed

| Step | Signal seen on November 9 | What it supports | Limit |
| --- | --- | --- | --- |
| 1 | `/process_api --firecracker-init`. Uptime about 19 minutes | A Firecracker microVM. Inferred: started that morning | One VM at one time. Uptime may not count time spent paused |
| 2 | The shell is a child of `/opt/claude-code/bin/claude`, launched by `environment-manager task-run --session-mode resume` | The harness is Claude Code, started for each session | It does not show what other surfaces use |
| 3 | The `claude` command on the PATH reports 2.1.319 | A Claude Code install is present | It is the same file as the running program, `/opt/claude-code/bin/claude` (checked with `readlink -f`) |
| 4 | `REMOTE=true`, `ENTRYPOINT=remote_cowork`, `SESSION_ORIGIN=claude_ai_chat`, `VERSION=2.1.42`, API base `api.anthropic.com` | A remote session started from chat. Model requests go to the configured endpoint | Variable names may change between releases |
| 5 | A `-home-claude` project folder | The transcript is kept in the VM | It is not the durable record |

### The two VMs, side by side

On November 16, 2026, Panaversity ran a second inspection, in a Claude conversation linked to a Mac with one connected folder. The cloud workspace uses the `Bash` tool. The desktop workspace uses the `device_bash` tool, which works only once a folder is connected. Ask Claude to run the same commands with each tool and show you the output.

```
# Identity, operating system and kernel
whoami; uname -a; cat /etc/os-release | head -3
# CPU, memory and disk
nproc; free -h; df -h $HOME
# Connected folders (desktop workspace only)
ls $HOME/mnt/
# The first running processes
ps aux | head -12
# Common tools
which python3 node git ffmpeg pandoc soffice
```

| Signal | Cloud workspace | Desktop workspace |
| --- | --- | --- |
| User and host | `root` on host `vm` | A per-session user, `rcw-` and the session id, on host `claude` |
| Operating system | Ubuntu 24.04, x86_64, kernel 6.18 built for Firecracker | Ubuntu 22.04.5, arm64 (aarch64), kernel 6.8 |
| CPU, memory, free disk | 2, 7.8 GiB, about 31 GB | 4, 3.8 GiB, about 8 GB |
| First process | `/process_api --firecracker-init` (seen on November 9) | `bwrap --new-session --unshare-net`, with `socat` relays to a local proxy |
| Tools found | Python, Node, Git, FFmpeg, LibreOffice and Pandoc, among others | Python 3.10, Node, Git, FFmpeg, Pandoc and LibreOffice |
| Network test | `pypi.org` answered with HTTP 200 | `pypi.org` answered with HTTP 200. A news site did not answer, under the **Package managers only** setting |

Three more signals came from the desktop workspace's own tools on the same day. Staging allowed at most 50 files, 400 MB per file and 500 MB per call. A commit was refused when the file on the computer had changed since it was staged. Deleting in the connected folder failed with "Operation not permitted" until deletion was approved for that folder, which also blocked git until then.

| Signal | What it supports | Limit |
| --- | --- | --- |
| Different OS, architecture and size | Two separate machines with no shared disk | One account, one Mac, one day |
| `bwrap` with `--unshare-net` as the first process | Each session's commands run in their own sandbox inside the local VM, with network only through a proxy | The VM's own design is described only by Anthropic's overview |
| Staging and commit limits | Files cross between the two only through named steps | Tool limits are not published and can change |

### Where results land

On November 17, 2026, we looked at how one Claude desktop session, on a paid plan with one connected folder, decides where a result goes. The instructions Claude was following in that session set this order: a folder you name, then a file format you name, then the folder whose files the job works on (the result saved next to them, as a new file), then an artifact of the matching kind for anything to keep, and a normal reply for a one-time question. Connecting a folder did not change this on its own. The same instructions named the usual formats when none is asked for: .docx for a document, .pptx for a presentation, .xlsx for a spreadsheet and PDF for a design. We also saw Claude edit .md and .csv files in the connected folder from the desk on your computer, without staging them to the cloud desk.

| Signal | What it supports | Limit |
| --- | --- | --- |
| The order in Claude's instructions | The order in [Where your work is saved](../../claude-and-chatgpt-2/build-something/where-your-work-is-saved.md) | One session, one day. A summary of instructions, not a test of every case. Anthropic does not document it |
| Usual formats in Claude's instructions | Claude's usual choices, not fixed rules | Not tested one by one. Name the format to be sure |
| Editing .md and .csv in place | Plain-text files need no trip to the cloud desk | Seen, not documented |

### A reset experiment

Run step 1 at the start of two turns, with an idle gap between them, and write down both times. In the first turn, also write a small marker file to the VM's own disk, for example `date > /tmp/marker.txt`.

- If the marker is still there in the second turn, the same VM was restored. [The sandbox and its state](../../claude-and-chatgpt-2/architecture/the-sandbox-and-its-state.md#when-the-vm-is-paused-restored-or-replaced) quotes the documented pause and restore.
- If the marker is gone, the VM was reclaimed and a fresh one started.
- Do not rely on uptime alone. A restored VM may not count the time it spent paused, so a short uptime does not prove a new machine.

On November 8 and 9, the uptime was short and the demo file came back only as a copy from the conversation. The marker test was not run then. How long a paused VM is kept is still unknown.

### Caveats

Claims in this section carry the labels Observed, Reported by ChatGPT, Tool-documented, Vendor-documented, Inferred or Unknown. The [overview](../../claude-and-chatgpt-2/architecture/overview.md) explains each one.

The observations come from one cloud Claude chat session on November 8 and 9, 2026, from one conversation linked to a Mac on November 16, 2026, and from one Claude desktop session on November 17, 2026. They may not hold for every plan, account or surface. Beyond Anthropic's published overview, the internal architecture cannot be seen from inside a session. Vendor behavior, feature flags and availability may change.

### Sources

The full details of each source are in this page's source list.

- Claude Desktop, Code tab docs (Anthropic): the environments Local, Cloud, SSH and WSL.
- Panaversity's inspection of a cloud Claude session: a Python test run on November 8, 2026, and a process and environment inspection on November 9, 2026. The method is on this page.
- Panaversity's inspection of both VMs in a Claude conversation linked to a Mac, November 16, 2026. The method is on this page.
- Panaversity's observation of where results land, in a Claude desktop session, November 17, 2026. What was seen is on this page.
- ChatGPT's inspection of its own Python and terminal environment, November 9, 2026, as reported to Panaversity.
- Agents API architecture (OpenAI API docs): the hosted harness, the environment and the application server as separate parts.
- ChatGPT Work cloud security (ChatGPT docs): Work Cloud's own execution environment.
- Claude Managed Agents overview (Claude Platform docs): the hosted harness and sandbox, sessions, events and beta status.
- Managed Better Auth is now an OAuth provider, with MCP support (Neon, November 2026).
- Better Auth MCP plugin (Better Auth docs).
- Schedule recurring tasks in Claude Code Desktop (Claude Code docs): local, cloud and `/loop` tasks, missed runs, permissions and `requiresUserInteraction`.
- Let Claude coordinate ongoing work with Projects (Claude Code docs): the coordinator, threads, standing context, sandbox resume, local threads, limits and beta status.
- Projects redesigned: from folder to conversation (Anthropic blog, September 17, 2026).
- Use Claude Code in the cloud (Claude Code docs): idle sessions, reclaimed VMs and what is restored.
- Configure cloud environments (Claude Code docs): idle VMs pausing and being restored, network access levels, setup scripts and the environment cache, environment variables, network secrets, the GitHub proxy and shared environments.
- Self-hosted environments (Claude Code docs): environments, runners and sessions, leases, what stays on your machines, model routing, availability and limits.
- Verify session identity in self-hosted environments (Claude Code docs): the session token, its claims, what it proves and how to check it.
- Message your other Claude Code sessions (Claude Code docs): how messages travel, what a message can and cannot do, inbound controls and limits.
- How agent identity works, and Security and data handling (Claude Tag docs): service accounts, the credential store, access scope, sandboxes for each thread and availability.
- Automate work with scheduled tasks (Claude Code docs): cloud runs as full Claude Code cloud sessions, fresh copies of repositories, auto mode and limits.
- Claude Cowork architecture overview (Anthropic Help Center): the sandbox for each session, the egress proxy, credentials and plan availability.
- Unlocking the Codex harness: how we built the App Server (OpenAI, February 2026).
- Introducing ChatGPT agent (OpenAI, July 2025).
- ChatGPT: What's new (OpenAI).
- ChatGPT sandbox docs (OpenAI).
- OpenAI is folding Codex into the ChatGPT app (The New Stack, July 2026).
- Work with Claude Tag, overview (Claude Tag docs): Claude in team Slack channels, plans and availability.
- Run agents in parallel (Claude Code docs): subagents, agent view, agent teams, dynamic workflows and projects compared.
- Linux Foundation launches the Agent2Agent protocol project (Linux Foundation, June 23, 2025).
- [Tour of the Apps](../../claude-and-chatgpt-2/tour/overview.md): the product features this section builds on.
