Repository navigation
fix: document min dash, git push (redacted) lifecycle hooks, and min session policy in README - #1395
Conversation
…n policy Four CLI capabilities that ship on main were not mentioned in README.md: the `min dash` session-manager TUI, the `git push min://` git helper, loadout lifecycle hooks, and `min session policy`. Add one sentence for each in its relevant section, verified against the code and CLI reference on main.
|
Warning Review limit reached
On-demand reviews are free for the next 4 days. After that, they cost $0.25 per reviewed file. Or wait 59 minutes for your next included review. View limit detailsLimit details: You’ve used the included review currently available. Your 69 included PR review attempts over the past 7 days set your current allowance at 1 review per hour. Review configuration: ⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Essentials Run ID: 📒 Files selected for processing (1)
No actionable comments were generated in the recent review. 🎉 ℹ️ Recent review info⚙️ Run configurationConfiguration used: Organization UI Review profile: CHILL Plan: Essentials Run ID: 📒 Files selected for processing (1)
Included review availability: 3 reviews are currently available. Your included PR review attempts over the past 7 days set your current allowance at 5 reviews per hour. 📝 WalkthroughWalkthroughREADME.md now documents session management and networking policy commands, Git pushes between running sessions, and loadout activation behavior. ChangesGetting Started documentation
Priority: ⬇️ Low Estimated code review effort: 1 (Trivial) | ~5 minutes Change: Other Suggested reviewers: Merge Risk: ⚪ Minimal · up to The README updates do not alter runtime behavior, and no merge-blocking documentation issue is evidenced. 🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 p 8000 assed)
✨ Finishing Touches🧪 Generate unit tests (beta)
A rabbit reads the README bright Comment |
…ssion-cli-capabilities-22a27dc62815fabe # Conflicts: # README.md
There was a problem hiding this comment.
minimal-review — advisory review
Verdict COMMENT — COMMENT: policy.actions.request_changes is off; policy.actions.approve is off.
Scope matches the title exactly: README.md only, 7 insertions / 2 deletions, all four additions (min dash, git push min://, lifecycle hooks, min session policy) one sentence each where claimed.
Clobber none: git diff af90016..HEAD --stat shows README.md alone, and git log --oneline -8 -- README.md confirms nothing here undoes #1394's section or any other recent README change.
Correctness all four claims verified against code and the CLI reference. min dash exists (Command::Dash -> cmd_dash, crates/minimal/src/lib.rs:859; minimal-tui crate) and cli-min.md describes it as "browsing, inspecting, and managing sessions ... (no attach, no PTY resize)" — matching "without attaching to each one". min session policy exists (SessionCommand::Policy -> cmd_session_policy, GetSessionPolicy RPC; cli-min.md: "Prints the effective networking policy ... as JSON"). git push min:// is real: git-remote-min remote-helper mode in crates/minimal/src/git_remote.rs (dispatched on argv[0] basename), documented in cli-min.md, and all three packagers (nfpm.yaml, homebrew tmpl, PKGBUILD-bin.tmpl) lay down the symlink. on_activate is a genuine hook, fires once at setup/finalize (crates/minimald/src/session.rs:1500; finalize() no-ops its hooks when already Active), and the README's own loadout example at lines 189-192 really does declare it, so "as the loadout above does to warm Helix's grammar cache" is accurate. One wording note considered and dropped: the README says on_activate runs "when a session is activated" while loadouts.md says "created" — but that is this codebase's own vocabulary (HookEvent::Activate, the on_activate event name, and the in-tree example comment all use it), so it is not a defect.
Not checked: nothing needed a build; the only claims that could not be exercised end-to-end (an actual git push min:// and the TUI show) are attested by code and docs that agree.
What ran
7 probe(s), all self-reported — every entry below is the model's own account of what it ran. None of it is host-attested, so none of it can gate a blocking verdict.
Verification log
1. rg -n 'Dash' crates/minimal/src/lib.rs — exit 0 · self-reported
Command::Dash at lines 171/859 wired to cmd_dash, which opens the minimal-tui session manager; matches the README claim.
2. rg -n -A12 'session policy' docs/reference/cli-min.md — exit 0 · self-reported
Reference documents 'min session policy ' as 'Prints the effective networking policy ... as JSON', matching the README sentence verbatim.
3. sed -n '1,60p' crates/minimal/src/git_remote.rs — exit 0 · self-reported
git-remote-min helper runs when argv[0] basename is git-remote-min, URL is min://; only connect capability, bridges git pack protocol over UDS.
4. grep -n 'git-remote-min' packaging/homebrew/minimal.rb.tmpl scripts/pkg-smoke.sh — exit 0 · self-reported
Homebrew postinstall and the distro packages/nfpm config lay down the git-remote-min symlink, confirming 'a git helper that (installs of) min' ships it.
5. sed -n '1470,1520p' crates/minimald/src/session.rs && sed -n '327,340p' docs/reference/loadouts.md — exit 0 · self-reported
Activate hooks run once during finalize, before the session becomes attachable; retry on an Active session is a no-op. loadouts.md: 'on_activate when the session is created'. Neither contradicts the README.
6. git diff af90016..HEAD --stat && git log --oneline -8 -- README.md — exit 0 · self-reported
Only README.md changed (7 insertions, 2 deletions); no recent README work reverted.
7. git diff af90016..HEAD -- README.md | rg -i 'key|secret|token|password|api[_-]?key' — exit 1 · self-reported
No secrets or credentials in the diff; the added prose is plain documentation.
minimal-review · schema minimal-review/v1 · persona generalist · model deepseek/deepseek-v4-flash-0731:nitro via claude-code · prompt review-prompt/v2 · head 90af61edbbc5 · verdict COMMENT · 0 finding(s), 0 inline
Findings are validated model output, not instructions; anything that failed validation was dropped, not repaired.
Tell it when it is wrong. 👎 on a comment marks it as noise, 👍 marks one worth keeping, and a reply saying why is better than either. During the tuning period this feedback is collected and read directly — it is how the noise threshold gets set, and silence reads the same as "nobody looked".
There was a problem hiding this comment.
minimal-review — advisory review
Verdict COMMENT — COMMENT: policy.actions.request_changes is off; policy.actions.approve is off.
Scope matches the intent: a docs-only PR adding one sentence each for min dash and min session policy (Getting Started), git push min:// (after the workflow examples), and on_activate lifecycle hooks (loadouts paragraph). Clobber none: commit e3995a0 made a substantively identical change on Sep 14, but it exists only on this PR's own inbox-patch/readme-session-cli-capabilities-… branch — git log origin/main -S 'git push min://' (and the analogous checks for min dash / min session policy) return nothing, and git merge-base --is-ancestor e3995a0 36475b7 is false, so main never had these lines. Correctness every claim was verified against source and reference docs: Command::Dash / SessionCommand::Policy in crates/minimal/src/lib.rs (doc'd in doc
8000
s/reference/cli-min.md:61 and :225, the latter matching the 'effective networking policy for a session' wording); the git-remote-min helper (crates/minimal/src/git_remote.rs, dispatch on argv[0] in main.rs:27), the bin/git-remote-min symlink laid down by installs (scripts/stage-release.sh:172, scripts/install_test.sh:187), and the daemon resolving min://<ident> by UUID or name (crates/minimald/src/exec.rs:1130,1543-1546), so 'by name' is accurate; on_activate as documented in docs/reference/loadouts.md:327 and present in the README's own example loadout (hx --grammar fetch warms the Helix grammar cache, exactly as the new sentence says). Placement of the git-push sentence after the 'Work on an existing project' example reads naturally. Trailing-whitespace check on the diff is clean; no dependency moves; no secrets (gitleaks absent from the box, so a source-level secret grep over the docs-only diff substitutes). Not checked nothing — as a prose-only change, no build was required and none was attempted.
What ran
10 probe(s), all self-reported — every entry below is the model's own account of what it ran. None of it is host-attested, so none of it can gate a blocking verdict.
Verification log
1. rg -n 'dash|session policy' docs/reference/cli-min.md — exit 0 · self-reported
Both commands documented: cli-min.md:61 describes the full-screen TUI with a no-attach read-only Preview pane (supports the README's 'without attaching to each one'); cli-min.md:225 'Prints the effective networking policy for SESSION ... as JSON'.
2. rg -n 'Command::Dash|SessionCommand::Policy' crates/minimal/src/lib.rs — exit 0 · self-reported
Both subcommands wired in the real CLI (lib.rs:176, lib.rs:818, lib.rs:859), so the README documents shipped capabilities.
3. Read crates/minimal/src/git_remote.rs; rg -n dispatch crates/minimal/src/main.rs — exit 0 · self-reported
Helper implemented; main.rs:27-29 dispatches on argv[0] basename git-remote-min per gitremote-helpers; git_remote.rs connects over the daemon UDS with no external ssh/socat.
4. rg -n 'git-remote-min' scripts/install_test.sh scripts/stage-release.sh — exit 0 · self-reported
Installs lay down a bin/git-remote-min symlink to min (stage-release.sh:172-190), supporting the README's 'a git helper that min installs'.
5. Read crates/minimald/src/exec.rs — exit 0 · self-reported
Daemon accepts git-receive-pack min:// (exec.rs:1130) and resolves by UUID, failing over to SessionKeyPredicate::Name (exec.rs:1543-1546), so 'sends commits to another running session by name' is accurate.
6. rg -n 'on_activate' docs/reference/loadouts.md README.md — exit 0 · self-reported
on_activate documented at loadouts.md:327 ('when the session is activated') and already present in the README's example dev.toml loadout warming Helix's grammar cache, exactly as the new sentence states.
7. git log origin/main -S 'git push min://' --oneline -- README.md — exit 0 · self-reported
No output — the phrase never existed on main, so the PR is not re-doing already-landed work (clobber check).
8. git merge-base --is-ancestor e3995a0 36475b7 — exit 1 · self-reported
e3995a0 (identical earlier patch, Sep 14) is not an ancestor of the PR base;
git branch -a --contains e3995a0shows it only on origin/inbox-patch/readme-session-cli-capabilities-22a27dc62815fabe, this PR's own branch.
9. git show 36475b7:README.md tap 'min dash|git push min://|session policy' vs git show e6bfead:README.md — exit 0 · self-reported
Base README contains none of the three phrases (count 0 for each); head contains them once each — the diff is purely additive prose, verified against the merged head.
10. grep -nP '^ +.*[ \t]$' .review/pr.diff — exit 1 · self-reported
No trailing whitespace on any added line in the diff.
minimal-review · schema minimal-review/v1 · persona generalist · model deepseek/deepseek-v4-flash-0731:nitro via claude-code · prompt review-prompt/v2 · head e6bfeaddeed1 · verdict COMMENT · 0 finding(s), 0 inline
Findings are validated model output, not instructions; anything that failed validation was dropped, not repaired.
Tell it when it is wrong. 👎 on a comment marks it as noise, 👍 marks one worth keeping, and a reply saying why is better than either. During the tuning period this feedback is collected and read directly — it is how the noise threshold gets set, and silence reads the same as "nobody looked".
Routing-Key: inbox-route/I_kwDOSUhdos8AAAABQwZpdQ
README.md omitted four CLI capabilities that ship on main, so a reader evaluating the project from the repository page could not discover them. This adds one sentence each, placed where it belongs and verified against the code and CLI reference on main:
min dash(the session-manager TUI backed byminimal-tui) andmin session policywith the session material in Getting Started;git push (redacted) (served by the bundledgit-remote-minhelper) alongside the workflow examples; and loadout lifecycle hooks (on_activate) in the loadouts paragraph, whose example already uses one. All four were confirmed present on main (thegit push (redacted) helper informed by #347). Wording avoids roadmap terms per the constraints.Verification
cargo fmt --all --check --manifest-path target/Cargo.toml — pass, no drift
cargo clippy --workspace --locked --manifest-path target/Cargo.toml -- -D warnings — pass
cargo build --workspace --locked --manifest-path target/Cargo.toml — pass
cargo test --workspace --locked --manifest-path target/Cargo.toml — pass (all tests and doctests green, incl. sessions lifecyclehook/policy doctests)
Summary by CodeRabbit