fix: Hide `-c`/`--command` on `min session attach` from the CLI help surface by gominimal-aw-bot[bot] · Pull Request #1071 · gominimal/minimal

gominimal-aw-bot · 2026-07-29T19:36:20Z

Fixes #1062
Routing-Key: inbox-route/I_kwDOSUhdos8AAAABKsZxwQ

AttachArgs::command in crates/minimal/src/lib.rs carried #[arg(long, short)] with no hide, so -c / --command appeared in min session attach --help, -h, and shell completions — advertising a general remote exec the flag cannot deliver (no PTY, only three daemon commands accepted). Adding hide = true removes it from the visible surface while keeping it functional for existing scripted callers. The one coupled judgment call: the non-TTY attach error in ensure_interactive_attach_tty still names --command; rather than drop the reference (which would strip the only forward path for non-TTY callers and break its existing test), I reworded it to frame the flag as a deliberate hidden escape hatch, keeping the advice actionable and self-contained.

Verification

cargo fmt --all --check --manifest-path target/Cargo.toml — clean, no drift
cargo clippy --workspace --locked -- -D warnings — 0 warnings, finished ok
cargo build --workspace --locked — exit 0
cargo test --workspace --locked — all suites passed (exit 0), incl. interactive_attach_requires_a_tty

Generated by inbox-patch · ◷

Note

Hide `-c`/`--command` flag from `min session attach` CLI help and remove `--force` from `min init`

Marks the --command flag on min session attach as hidden in lib.rs so it no longer appears in help output, but remains functional for scripted callers.
Removes the --force flag from min init; run_init_flow now always overwrites minimal.toml if it exists, requiring only the --yes flag for non-interactive use.
Removes SHELL=/bin/sh override and SendEnv forwarding of LANG/LC_*/TZ from the SSH attach helper; only MINIMAL_SESSION_ID continues to be forwarded.
Behavioral Change: min init will overwrite an existing minimal.toml without --force; any callers relying on that guard will no longer be protected.

^{Macroscope summarized c0f9a4f. (Automatic summaries will resume when PR exits draft mode or review begins).}

Summary by CodeRabbit

Bug Fixes
- Improved the guidance shown when session attachment is attempted without an interactive terminal.
- Clarified the scripted attachment option using the hidden --command flag.
CLI Improvements
- The --command option is now hidden from standard help output while remaining available for scripted use.

Add `hide = true` to the `#[arg(long, short)]` attribute on `AttachArgs::command` so `-c` / `--command` no longer appear in `min session attach --help`, `-h`, or shell completions. The flag stays functional for existing scripted callers; the option promises a general remote exec it cannot deliver (no PTY, only three daemon commands accepted), so it should not be advertised for discovery. The non-TTY attach error still points stuck callers at `--command`; reword it to frame the flag as a deliberate hidden escape hatch rather than drop the reference, keeping the advice actionable now that the flag is off the help surface.

coderabbitai · 2026-07-29T19:36:28Z

📝 Walkthrough

Walkthrough

The min session attach command flag is hidden from Clap help output, while non-TTY error guidance continues to identify --command <cmd> for scripted callers.

Changes

Attach CLI surface

Layer / File(s)	Summary
Hide command flag and update non-TTY guidance `crates/minimal/src/lib.rs`	`AttachArgs.command` now hides `-c` and `--command` from help output, and `ensure_interactive_attach_tty` describes the hidden scripted escape hatch.

Estimated code review effort: 1 (Trivial) | ~5 minutes

Possibly related issues

gominimal/inbox#416 — Directly covers hiding AttachArgs::command and updating its non-TTY guidance.

Possibly related PRs

gominimal/minimal#956 — Adds the related non-TTY fail-fast behavior for interactive attach.
gominimal/minimal#1036 — Aligns non-interactive attach refusal behavior in the same source file.
gominimal/minimal#846 — Makes related Clap argument visibility changes for hidden CLI flags.

Suggested reviewers: norrietaylor, evanspearman, twitchyliquid64

Poem

A bunny tucked the command from sight,
While keeping its scripted path alight.
“Use the hidden flag,” errors say,
“When terminals hop away.”
CLI help now feels just right.

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Linked Issues check	✅ Passed	The change satisfies `#1062` by hiding `-c/--command` in clap while preserving the non-TTY hint and existing functionality.
Out of Scope Changes check	✅ Passed	The final diff only touches the attach CLI surface and error message, with no unrelated code changes.
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%.
Title check	✅ Passed	The title clearly and concisely describes the main change: hiding the attach command flag from help output.
Description check	✅ Passed	The description includes a clear summary and verification steps, but it does not follow the template headings exactly and omits the checklist.

✨ Finishing Touches

📝 Generate docstrings

Create stacked PR
Commit on current branch

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

…d4710eb24

The merge commit c0f9a4f resolved conflicts in crates/minimal/src/lib.rs by taking the pre-merge side, silently reverting two fixes that had landed on main: - #1053 — the `min init` refuse-to-overwrite guard: `InitArgs::force`, the bail on an existing `minimal.toml`, and the Updated/Created wording. Without it `min init` overwrites an existing `minimal.toml` with no backup and no prompt, which is the data loss #1032 was filed for. - #1060 — the `SHELL=/bin/sh` pin on the ssh ProxyCommand and the `SendEnv` forwarding of LANG/LC_*/TZ. Without the pin, attach dies at "banner exchange ... Broken pipe" for any user whose $SHELL is a bare name (fish) or absent from the ssh context. Neither revert was caught by the suite: there is no test for the init guard, and `interactive_attach_requires_a_tty_on_stdin` only asserts the error contains "not a TTY" and "--command", which holds either way. Rebuilt lib.rs from main and reapplied only the intended change, so the diff against main is now exactly the `hide = true` on `AttachArgs::command` plus the non-TTY error reword: +4/-3, was +9/-51. Verified: rustfmt clean; `cargo build -p minimal --locked` ok; `min session attach --help` no longer lists `-c`/`--command` while `--command` still parses; `min init --help` lists `--force` again. `cargo test -p minimal` cannot run on macOS (dev-deps pull minimald -> procfs/caps, Linux-only), so the suite is left to CI. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>

norrietaylor · 2026-07-29T21:06:02Z

Corrected: the merge commit was reverting two fixes that landed on `main`

Pushed d681e4d. The diff against main is now +4/-3 — it was +9/-51.

What was wrong

c0f9a4f ("Merge branch 'main' into inbox-patch/…") resolved the conflict in crates/minimal/src/lib.rs by taking the pre-merge side of the whole file. That silently reverted two commits that had landed on main in the meantime:

Reverted	What it protected
`f36045f9` #1053	`InitArgs::force`, the bail on an existing `minimal.toml`, and the Updated/Created wording
`12c81c3f` #1060	the `SHELL=/bin/sh` pin on the ssh ProxyCommand, and `SendEnv` forwarding of `LANG`/`LC_*`/`TZ`

Consequences had it merged:

min init would overwrite an existing minimal.toml with no backup and no prompt — exactly the data loss min init destroys an existing minimal.toml with no backup; --no-input is never read and confirm() treats EOF as yes #1032 was filed for and fix: refuse to overwrite existing minimal.toml without --force #1053 fixed, reintroduced four commits later.
min session attach would break for any user whose $SHELL is a bare name (fish) or absent from the ssh context. Per feat: always set certain important variables automatically #1060's own comment, ssh execs $SHELL with no PATH lookup, so the transport dies at banner exchange … Broken pipe.
Locale and timezone would stop reaching the session.

Verified by fetching the file at both ends of the PR:

MARKER (from #1053 / #1060)                base     HEAD(before fix)
refusing to overwrite existing             1        0
ssh.env("SHELL", "/bin/sh")                1        0
SendEnv=LANG                               1        0
SendEnv=LC_*                               1        0
SendEnv=TZ                                 1        0

Why the green suite did not catch it

The PR body reports cargo test --workspace --locked passing, and that is consistent — no test covers either reverted behaviour:

There is no test for the min init overwrite guard.
interactive_attach_requires_a_tty_on_stdin asserts only err.contains("not a TTY") && err.contains("--command"), which holds for the old message, the reworded one, and the reverted one.

So the suite was never going to fail here. Worth adding a test for the init guard, given it has now been broken once and silently re-broken once.

What I changed

Rebuilt crates/minimal/src/lib.rs from origin/main (byte-identical, confirmed by an empty diff) and reapplied only the intended change:

hide = true on AttachArgs::command — the actual fix for Hide min session attach -c/--command from the CLI help surface #1062.
The non-TTY error reword, kept as authored. It stays within Hide min session attach -c/--command from the CLI help surface #1062's scope note ("keep the option functional") and the existing test still passes.

Verification

rustfmt --check on the changed file — clean
cargo build -p minimal --locked — Finished dev profile in 3m05s
min session attach --help — no -c/--command line; min session attach --command still errors with a value is required for '--command <COMMAND>', so the flag remains functional
min init --help — --force is listed again
cargo test -p minimal cannot run on macOS (dev-deps pull minimald → procfs/caps, both Linux-only, exactly as AGENTS.md documents). The suite is left to CI.

For the `inbox-patch` workflow

This is a merge-conflict-resolution failure, not a bad patch — c4231296, the generated commit, was correct on its own. The failure mode is worth guarding: when the bot merges main into a patch branch and the target file has moved, whole-file resolution reverts everything that landed in between, and the PR body still describes only the intended change. A git diff origin/main --stat sanity check against the expected footprint would have caught it.

hide = true on AttachArgs::command (#1071) was silently reverted by commit 64a0764, an unrelated help-text reformat -- the same collateral-clobber pattern #1071 itself had to repair once already. Restore it, along with the "hidden escape hatch" non-TTY error wording that shipped with it. This PR went the opposite direction from a hidden flag: it documented the full min run/min package build/min check allowlist in docs/concepts/sessions.md, docs/guide/agents.md, and docs/reference/cli-min.md. Since -c/--command is a hidden legacy escape hatch and not a general remote shell, remove those mentions instead of expanding them. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

…cs (#1110) * docs: name full exec-channel allowlist for session attach -c Three docs stated that the non-interactive `min session attach -c` channel accepts only `min run` invocations. The daemon's exec handler also accepts `min package build [args...]` and `min check [args...]` (crates/minimald/src/exec.rs), so the docs understated the allowlist. List all three accepted forms in cli-min.md, agents.md, and sessions.md. Fixes #1067 * chore: no-op to restart ci * fix: keep session attach -c hidden, restore the dropped hide=true hide = true on AttachArgs::command (#1071) was silently reverted by commit 64a0764, an unrelated help-text reformat -- the same collateral-clobber pattern #1071 itself had to repair once already. Restore it, along with the "hidden escape hatch" non-TTY error wording that shipped with it. This PR went the opposite direction from a hidden flag: it documented the full min run/min package build/min check allowlist in docs/concepts/sessions.md, docs/guide/agents.md, and docs/reference/cli-min.md. Since -c/--command is a hidden legacy escape hatch and not a general remote shell, remove those mentions instead of expanding them. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * docs: stop enumerating the -c/--command allowlist in AGENTS.md too The footgun note still spelled out the exact min run/min package build/min check allowlist, which is itself advertising a hidden flag. Keep the warning (don't treat -c as a general shell) and the source pointer, drop the enumerated command list. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> * fix: stop pointing the non-TTY attach error at --command The non-TTY error was still advertising the hidden --command escape hatch to every interactive caller who hit it -- inconsistent with hiding the flag everywhere else. Drop the mention; it only ever fires on the interactive path (command.is_none()), so it was pushing newcomers toward a flag we deliberately don't document. Update the one test that asserted on the old wording. Also drop the AGENTS.md footgun bullet entirely per feedback -- no partial version, just remove it. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com> --------- Co-authored-by: gominimal-aw-bot[bot] <281738952+gominimal-aw-bot[bot]@users.noreply.github.com> Co-authored-by: bryan <bryan@minimal.dev> Co-authored-by: Norrie Taylor <norrie@minimal.dev> Co-authored-by: Claude Sonnet 5 <noreply@anthropic.com>

gominimal-aw-bot Bot mentioned this pull request Jul 29, 2026

Hide min session attach -c/--command from the CLI help surface #1062

Closed

norrietaylor and others added 2 commits July 29, 2026 12:41

Merge branch 'main' into inbox-patch/hide-attach-command-flag-532408b…

c0f9a4f

…d4710eb24

twitchyliquid64 approved these changes Jul 29, 2026

View reviewed changes

twitchyliquid64 marked this pull request as ready for review July 29, 2026 21:06

norrietaylor enabled auto-merge (squash) July 29, 2026 21:09

norrietaylor merged commit 0e25232 into main Jul 29, 2026
29 checks passed

norrietaylor deleted the inbox-patch/hide-attach-command-flag-532408bd4710eb24 branch July 29, 2026 21:21

coderabbitai Bot mentioned this pull request Jul 29, 2026

fix: Give global CLI flags a short first help line #1084

Merged

coderabbitai Bot mentioned this pull request Aug 7, 2026

feat(min): move session exec to own subcommand #1186

Merged

Navigation Menu

Uh oh!

fix: Hide `-c`/`--command` on `min session attach` from the CLI help surface - #1071

fix: Hide `-c`/`--command` on `min session attach` from the CLI help surface#1071
norrietaylor merged 3 commits into
mainfrom
inbox-patch/hide-attach-command-flag-532408bd4710eb24

gominimal-aw-bot Bot commented Jul 29, 2026 •

edited by macroscopeapp Bot

Loading

Uh oh!

coderabbitai Bot commented Jul 29, 2026 •

edited

Loading

Walkthrough

Changes

Uh oh!

norrietaylor commented Jul 29, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Uh oh!

Conversation

gominimal-aw-bot Bot commented Jul 29, 2026 • edited by macroscopeapp Bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Verification

Hide -c/--command flag from min session attach CLI help and remove --force from min init

Summary by CodeRabbit

Uh oh!

coderabbitai Bot commented Jul 29, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Walkthrough

Changes

Uh oh!

norrietaylor commented Jul 29, 2026

Corrected: the merge commit was reverting two fixes that landed on main

What was wrong

Why the green suite did not catch it

What I changed

Verification

For the inbox-patch workflow

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

gominimal-aw-bot Bot commented Jul 29, 2026 •

edited by macroscopeapp Bot

Loading

Hide `-c`/`--command` flag from `min session attach` CLI help and remove `--force` from `min init`

coderabbitai Bot commented Jul 29, 2026 •

edited

Loading

Corrected: the merge commit was reverting two fixes that landed on `main`

For the `inbox-patch` workflow