Chatgpt Pro Line

ChatGPT Pro Line

Intended Use

Use ChatGPT Pro for higher-level intelligence tasks where a second strong model can materially improve the work:

• system architecture and technical strategy
• product and feature design decisions
• specification drafting and critique
• research synthesis, including online research when appropriate
• risk analysis, tradeoff review, and planning across files/repos/agents
• debugging strategy for confusing failures after Codex has gathered evidence

Do not burn the line on trivial syntax, mechanical edits, or proof-only smoke tests except when validating the transport itself. Transport tests prove the line works; they are not the product.

Ask for high-leverage judgment, not merely the smallest next patch. It is fine to request the most intelligent or bold implementation direction given the evidence; Codex can still execute that direction in small, testable slices. Invite ChatGPT Pro to disagree, ask clarifying questions, propose a different architecture, or recommend larger prerequisite work when that is the professional answer. The line is for senior planning and system judgment, not for rubber-stamping the current implementation.

Canonical Command

When this skill is installed from the Codex plugin, resolve the package root as two directories above this SKILL.md file and prefer the packaged CLI at:

<plugin-root>/bin/chatgpt-pro

Use chatgpt-pro directly only when that command is already on PATH. Inside this source repo, ./bin/chatgpt-pro ... and npm run chatgpt:call -- ... are development entrypoints for the same behavior.

First-run package flow:

chatgpt-pro init
chatgpt-pro doctor --warm
# complete visible login if prompted
chatgpt-pro doctor --live
chatgpt-pro status --alias=main

doctor --warm opens or reuses the dedicated visible ChatGPT browser profile without sending a prompt. doctor --live verifies login, composer, model, and intelligence state without sending a prompt, and refreshes the model cache.

chatgpt-pro call \
  --alias=main \
  --prompt-file=prompt.md \
  --upload-file=.devspace/context/current/repo-context.md \
  --response-mode=blocking

Equivalent minimal form:

chatgpt-pro call --alias=main --prompt="..."

Before a call, agents may inspect repo-local state without touching the browser:

chatgpt-pro status --alias=main

status is a registry/local command. It reports the current project id, room binding, latest receipt/transcript paths, browser/profile lock owner if busy, project-state lock owner if busy, and model-cache state. It must remain safe to run while another agent holds the live browser lock.

By default, chatgpt-pro call selects the live Pro intelligence level when the website exposes it. Use --level=... / --intelligence=... or CHATGPT_LEVEL to choose another live label. Use --no-default-pro only for transport debugging where changing the selector would obscure the failure.

Required Live Thread Output

When using this line in a Codex conversation, repeat the exchange verbatim in the thread using these exact headings:

## Message Sent To ChatGPT Pro

...

## Message Received From ChatGPT Pro

...

The run artifact transcript.md must use the same headings. The canonical renderer is src/transcript.mjs.

chatgpt-pro call prints this exact thread-echo block by default after the receipt paths. Copy the printed block into the Codex thread. Do not summarize, paraphrase, trim, or rewrite the Message Sent To ChatGPT Pro or Message Received From ChatGPT Pro sections. Additional commentary may come before or after the exact block, but not inside it.

When files are uploaded, the Message Sent To ChatGPT Pro block is the exact text typed into the composer. The uploaded file bodies are not pasted into the Codex thread; their paths, bytes, hashes, and upload status belong in the receipt.

chatgpt-pro read prints the exact received message. Machine callers may set CHATGPT_THREAD_ECHO=0, but interactive Codex use should leave echoing on and receipts must record threadEcho.mode, threadEcho.transcriptSha256, threadEcho.sentSha256, and threadEcho.receivedSha256.

Verify transcript integrity when needed:

chatgpt-pro transcript verify --receipt=.devspace/runs/<run-id>/receipt.json

Response Mode

Default and only implemented mode:

• blocking: send the prompt, wait for the anchored assistant response, write artifacts, then continue.

Reserved for later:

• event: send and return a run id immediately, then read/poll later.

If a caller requests anything except blocking, the command should fail rather than pretending event mode exists.

Context

Default posture: keep the composer text small and attach bulky context deliberately. chatgpt-pro call defaults to --repo-context=auto: it detects prompts that clearly ask for repo architecture, whole-codebase context, security review, or similar cross-file reasoning. Generated repo context is never uploaded or inlined unless the call also includes --confirm-repo-context-upload or CHATGPT_CONFIRM_REPO_CONTEXT_UPLOAD=1. It skips repo upload for trivial transport checks and when the caller already supplied --context-dir, --context-file, or --upload-file.

Use --repo-context=upload --confirm-repo-context-upload when broad repo context is definitely needed. Use --repo-context=inline --confirm-repo-context-upload only for tiny/debug contexts. Use --no-repo-context or --repo-context=off for narrow questions. The bundle lives under .devspace/context-bundles/<id>/ and includes:

• repo-context.md: one markdown file containing repo structure, LOC, file contents, and a source map
• manifest.json: machine-readable file and directory records, including line ranges in repo-context.md
• repo-context.zip: zipped repo-context.md plus manifest.json

Generate it directly when needed:

chatgpt-pro context bundle
npm run context:bundle

Attach additional files directly:

chatgpt-pro call --alias=main --upload-file=notes.md --upload-file=diff.patch --prompt="Review the attached artifacts."
CHATGPT_UPLOAD_FILES="notes.md,diff.patch" chatgpt-pro call --alias=main --prompt="Review the attached artifacts."

Repo context modes:

chatgpt-pro call --alias=main --repo-context=auto --confirm-repo-context-upload --prompt="Review the repo architecture."
chatgpt-pro call --alias=main --repo-context=upload --confirm-repo-context-upload --prompt="Use the attached repo context."
chatgpt-pro call --alias=main --repo-context=inline --confirm-repo-context-upload --prompt="Tiny contexts only."
chatgpt-pro call --alias=main --no-repo-context --prompt="No repo context."

Generated repo context has a fail-closed safety gate. It rejects secret-like paths and content such as .env*, .npmrc, .ssh/, .aws/, private keys, credentials files, known token formats, high-entropy secret assignments, symlinks, and realpath escapes before writing or uploading the bundle.

Use --context-dir only for small, deliberate text envelopes. Recognized files:

• codex-session-digest.md
• chatgpt-history.md
• repo-context.md, or manifest.json when repo context is absent
• diff.patch
• test-output.txt

The wrapper composes input.md, records inline context hashes/budgets, attachment paths/sizes/hashes/upload status in receipt.json, and saves the exact typed exchange in transcript.md.

For human-driven ChatGPT threads, bind the thread URL to a repo alias and export visible history before asking Codex to continue from it. Prefer rooms rebind because it binds or verifies the room without sending a prompt.

chatgpt-pro rooms rebind --alias=spec --conversation-url=https://chatgpt.com/c/...
chatgpt-pro history export --alias=spec --last=20
chatgpt-pro history export --alias=spec
chatgpt-pro history read --path=.devspace/chatgpt-history/spec/<run-id>/history.json

The history export writes chatgpt-history.md and history.json under .devspace/chatgpt-history/<alias>/..., including both user and assistant messages. Use --last=N for a recent window or omit it for the all-visible conversation snapshot. All-visible means all conversation turns currently loaded in the ChatGPT page DOM, not guaranteed account-complete history. History artifacts include historyCompleteness.claim, loadAllAttempted, absoluteConversationComplete, and olderMessagesMayExist.

For v1, history export --load-all must fail with mode.unsupported rather than pretending to scroll/load the whole account-side conversation. Use --require-visible-min=N when a caller needs to fail closed if too little history is loaded.

Repo-Scoped Rooms

Aliases are owned by the current repo. Before using a room alias, the wrapper checks .devspace/state/chatgpt-project.json, the git-level chatgpt-pro.projectId, and the canonical session registry:

~/.chatgpt-pro-codex/projects/<projectId>/chatgpt-sessions.json

Linked git worktrees share the same project id and therefore the same rooms. Separate git repositories get separate project ids. Non-git folders fall back to a realpath-derived project id.

Room policy:

• main: long-lived repo collaboration and implementation planning.
• debug: focused failure investigation while the bug family is the same.
• critic: independent review; prefer --fresh or --new for clean critique.
• scratch: disposable prompt and context checks.

Use:

chatgpt-pro rooms project
chatgpt-pro rooms list
chatgpt-pro rooms show --alias=main
chatgpt-pro rooms new --alias=critic
chatgpt-pro rooms rebind --alias=spec --conversation-url=https://chatgpt.com/c/...
chatgpt-pro rooms rebind --alias=spec --conversation-url=https://chatgpt.com/c/... --registry-only
chatgpt-pro rooms repair --alias=main

rooms project, rooms list, and rooms show are registry-only and may run while another agent holds the global browser lock for a live call. rooms new, default rooms rebind, and rooms repair are live-browser operations: they acquire the global browser lock, verify/open the target, mutate room lineage, and send no prompt. rooms rebind --registry-only is the explicit offline escape hatch; it validates the ChatGPT conversation URL shape but records roomTargetVerification: not_verified_registry_only until a later live operation verifies or repairs the target.

Thread modes:

chatgpt-pro call --alias=main --prompt="..."       # continue repo-owned main
chatgpt-pro call --alias=critic --fresh --prompt="..." # new unbound thread, alias hint only
chatgpt-pro call --alias=critic --new --prompt="..."   # new thread bound to critic
chatgpt-pro call --alias=main --rebind-alias --conversation-url=https://chatgpt.com/c/... --prompt="..."

--fresh must not mutate the alias. --new requires an alias and moves that alias to the newly opened thread while preserving lineage. --rebind-alias points an alias at an already-open conversation. Prefer the rooms commands when you need to create, rebind, or repair room state without sending a prompt; the call flags are send-and-mutate shortcuts.

For normal multi-agent use, do not open a new tab on every call. Create or bind a repo-owned room once, then keep using that alias:

chatgpt-pro rooms new --alias=critic
chatgpt-pro call --alias=critic --prompt-file=next-review.md

Use --fresh for clean one-off reviews where you explicitly do not want to change the alias. Use --new when the alias should move to a new long-lived thread.

If an alias belongs to another repo, fail closed with session.alias_project_mismatch. Rebind only when deliberate:

chatgpt-pro call --alias=main --rebind-alias --conversation-url=https://chatgpt.com/c/... --prompt="Deliberately rebind this repo room."
chatgpt-pro call --alias=main --rebind-alias --prompt="..."

Before every alias call or read, the wrapper must verify the browser target is showing the alias's expected ChatGPT conversation URL. If the saved target id is stale or points at a different thread, repair by finding an already-open target with the expected conversation id or by opening the expected URL. Do not send or read from a target that is not bound to the room; receipts should expose roomTarget.targetBoundToRoom.

Deterministic repo-room proof expected by this package:

• a base git repo and linked git worktree share projectId, canonical registry path, and room aliases
• a separate git repo gets a different projectId, registry path, and aliases
• a repo alias can deliberately move to a different ChatGPT thread while preserving archived lineage for the prior thread
• chatgpt-pro rooms list/show can inspect registries while the global browser lock is held by another operation
• chatgpt-pro history export verifies the alias target before reading, writes an all-visible or last-N history artifact with user and assistant messages, and chatgpt-pro history read can read it later without touching the browser

Live repo/thread proof expected before claiming the line is packaged:

• two separate fixture git repos can use the same alias name without sharing projectId, registry path, or ChatGPT thread
• a linked worktree can inspect and export the base repo's bound room history
• continuing a repo-owned alias reuses the existing ChatGPT conversation URL instead of opening a fresh thread by default
• a second alias in the same repo binds to a separate ChatGPT conversation URL
• all-visible history exports for each room read back from disk and contain the expected markers without leaking markers from another repo or alias

Run:

npm run live:repo-thread-matrix