[clintecker]

Problem

Last of three PRs for #732. PRs #733 and #734 can pause a gate and answer it from the library, but the CLI still reads gates from stdin without a terminal. An agent's shell, CI or a script therefore can't drive a gated workflow such as quickfix.

Fix

• --gates auto|console|pause.
  • With no gate automation flag, a run with no terminal on stdin, or with --json (agent harnesses may allocate a PTY), now pauses at a human gate.
  • console is the old stdin prompt; pause always pauses and implies --no-tui.
  • console / pause combined with --auto-approve, --webhook-url or --autopilot is a usage error.
• Exit 4 for paused_gate. paused_billing stays at 1; exitCodeFor centralizes the mapping. The summary prints the waiting gate (label, options, note, full prompt) with no failure banner and no tracker -r hint, and ends with the exact tracker answer command: -w / --artifact-dir repeated, --node when several gates wait.
• Invocation record in the secure state dir (invocation.json): argv minus -r / --from, cwd, and pipeline.GraphDigest of the loaded workflow and subgraphs. Sidecars are resolved into node attrs, so editing one changes the digest.
• tracker answer.
  • With no run ID it lists waiting runs; with a run ID it shows the gate.
  • With an answer — a label, --text, text after --, --default, --answers-file, --cancel, --node — it validates against the checkpoint record, rebuilds the run from the invocation record, refuses a changed workflow, re-checks the gate's nonce once the engine holds the run lock, and resumes in plain mode with the answer held in memory.
• Docs. cli.html (command, --gates, exit code 4), CLI help, README (new "Pausing at human gates" section), CHANGELOG (Added, plus Changed: no-terminal gates pause; --gates console restores stdin), glossary, architecture status table, AGENTS.md / llms.txt (what an agent does on exit 4), handlers/human.md, CLAUDE.md.

Verification

• Unit: gateModeFor table, --gates conflicts, interviewer selection (the old "non-TTY → console" case became "non-TTY → pause" plus "--gates console → console"), pause implies --no-tui, interpretRunResult / exitCodeFor, paused summary (last line = answer command), answer-command shapes, invocation round trip, stripResumeArgs, GraphDigest, answer argument parsing (incl. -- text), answers-file schema, list/show.
• End to end with the built binary and the fake claude:
  • quickfix pause → revise → --text → approve, committing to tracker/quickfix/<id>. --param change reached both MakeChange prompts and the checkout is untouched.
  • --default rejects; refusals (typo, bad answers file, unknown --node); list.
  • Digest refusal after editing the .dip (mutation-checked).
  • --json stream with exactly one terminal event (gate_paused) and no failure events.
  • Console -r resume after a pause, after which tracker answer reports "not waiting".
• go test ./... green; make doctor all A; make dippin-test green; golden traces unchanged; make docs-check (16 commands) green.

Stacked on #734. I'll rebase onto main once #733 and #734 merge.

Closes #732

🤖 Generated with Claude Code

---

^{Need help on this PR? Tag @codesmith-bot with what you need.}

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	⚠️ Failed 2026-10-20T23:17:17.785011Z We couldn’t complete this review. Comment @codex review to try again.	ad71b20	PR opened

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on base/target branches other than the default branch.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

• Configuration used: Organization UI
• Review profile: CHILL
• Plan: Advanced
• Run ID: 8b33ce7b-b27c-43ed-964b-729e51539799

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}