Files
gstack/ship/SKILL.md.tmpl
f67c478f05 v1.91.34.0 fix: /autoplan on current Claude Code and long sessions, quieter pre-push and sync guards, gstack-post, gstack-doctor (#3069)
* docs: Oct 7 follow-up wave plan (pre-review)

* docs: Oct 7 follow-up wave plan, reviewed through /autoplan

* docs: Oct 7 wave plan, record final gate decisions (UC1 accepted, UC2 rejected)

* docs: credit #1822 to @Mike-E-Log and #1743 to @habassa5 (v1.64.0.0)

The v1.64.0.0 CHANGELOG entry credited the extension token fix (#1822)
and the polyfill exited promise (#1743) to the wrong handle, and the
extension-sender-auth test header repeated the first slip. Commit history
is not rewritten.

Fixes #3054.

Co-authored-by: Mike Ilog <48659643+Mike-E-Log@users.noreply.github.com>
Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* fix(context-restore): offer newer checkpoints from nested repos and task worktrees

/context-restore could present an older checkpoint as the latest with no
error in two layouts:

1. A session started in a non-git parent saved from a nested repo, so the
   checkpoint went to the nested repo's bucket. Restore in the parent now
   lists checkpoints of other buckets whose project_root lies below the cwd
   (BELOW_CWD; the 200 newest are read, and from $HOME only each project's
   newest). /context-save also leaves a JSON pointer (checkpoint path, bucket,
   project_root) in the starting directory's bucket when $CLAUDE_PROJECT_DIR
   is set and files elsewhere; restore lists those as POINTER, in or out of git.
2. A plan checkpoint on the main tree's branch shadowed a newer checkpoint
   saved in a ticket worktree. When another branch holds a newer checkpoint
   for the same task (a worktree below the main tree, read from the new
   worktree: stamp; equal normalized titles; or a shared #123 / ABC-123
   token in titles or branch names), restore prints NEWER_TASK and the skill
   shows both and proposes the newer one. Generic titles (wip, checkpoint,
   save, notes, todo) never match on title alone, and the worktree rule
   applies only when restoring from the main tree, so sibling worktrees
   still never shadow each other.

The branch-first ordering itself is unchanged.

Refs #3065.

Co-authored-by: Thomas Gernhardt <t.gernhardt@softwareprojekt.net>
Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* fix(redact-prepush): list each MEDIUM finding with rule, file line and fix

The pre-push hook printed only a MEDIUM count, which nobody could review
without re-running the engine by hand. It now prints one line per finding:
rule id, file and file line, and a fix, never the matched value.

Line numbers come from the hunk headers (combined --cc headers included), so
a finding in a later hunk or past a scan-chunk boundary points at the real
file line. Findings from the per-commit catch-up path name the commit whose
file line it is, so content removed before the pushed tip is still locatable.

Plan items: CEO-7 (summary half), DX-9 (fix lines), ENG-8 (line mapping).

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* fix(redact-prepush): stop flagging your own email and already-public authors (#3060)

The hook never passed the engine's selfEmail / repoPublicEmails, so every
push that added the pusher's address, or a commit author's, raised a MEDIUM
pii.email warning for an address the destination already shows.

The hook now reads `git config user.email` and one bounded `git log` of
author and committer addresses, raw and mailmapped (%ae %aE %ce %cE): the
history of every pushed tip (those commits become public with the push),
each remote tip being replaced, and, for a new branch or an unknown remote
tip pushed to a configured remote, that remote's own tracking refs. A URL or
unknown destination never borrows tracking refs, and another remote's
history never counts.

The read stops at 50,000 commits (the newest authors are used) and 5 seconds.
An error or the timeout passes no history and, when an email finding is
listed, prints one line saying suppression was limited for this push. The two
inputs fail independently. HIGH findings and every other rule are untouched.

Controls: a stranger's address, an address only another remote knows, a URL
push next to a matching remote, and a first push of a repo with no history
all still report.

Plan items: A3, CEO-7, CEO-23 (A3 half), DX-9 (partial history), ENG-8
(history scope, mailmap).

Co-authored-by: akmandhania <143462686+akmandhania@users.noreply.github.com>
Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* feat(redact-prepush): per-address allowlist for pii.email (gstack.redact.allowEmail)

`git config --add gstack.redact.allowEmail <address>` allows that address for
MEDIUM pii.email findings only, so a team alias or support address that has
to be in the repo stops warning without reaching for --no-verify. It never
affects HIGH findings or any other rule, and the pii.email fix line names it.

Controls: a different address still reports, and a HIGH secret next to an
allowed address still blocks.

Plan item: DX-17.

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* docs(troubleshooting): pre-push pii.email findings, exemptions and allowlist

Plan item: DX-12 (pre-push touchpoint).

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* test(redact-prepush): assemble fixture addresses at runtime

The email fixtures were literals, so this branch's own push listed 47
pii.email findings in its test file. Build them from a helper instead, the
way the neighbouring tests build credential and IP fixtures.

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* fix(browse): offer the user's signed-in tab, name silent aside repl failures

Rules (rendered into every browsing skill, kept short):
- Rule 1: before the first openTab, offer the user's tabs on the target
  origin (title and origin only; the rest of listBrowserTabs() stays
  private) and attach only after the user confirms one.
- Rule 4: a second sign-in wall after a confirmed sign-in means the session
  is bound to the tab or URL; offer that tab instead of another sign-in.
- Rule 6 no longer suggests an early return; rule 7 says a fast [ok without
  GSTACK_STEP_OK is an abort.

Cookbook (browse, devex-review): a "Silent failures" list (evaluate returns
only serializable values, no top-level return, locator(sel).press(key), an
empty DOM read is about the selector until the screenshot and the request's
response say otherwise, read toggle state before clicking) and a signed-in
tab recipe that prints only target-origin tabs and never closes the user's tab.

Parity caps for design-consultation, land-and-deploy, qa and the browse
skeleton are raised to the measured sizes, with notes.

Refs #3063, #3064.

Co-authored-by: Thomas Gernhardt <t.gernhardt@softwareprojekt.net>
Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* feat(codex): run gstack-codex-probe as a command, one subcommand per probe

Generated skills now execute bin/gstack-codex-probe (select-model, check-auth,
show-sandbox, check-sandbox, probe-model, check-version, show-first-use-notice,
run-with-timeout, log-event, log-hang, help) and capture its KEY: value lines,
so no helper state crosses into the calling shell and the preflight works from
bash, zsh or sh. select-model keeps --model > GSTACK_CODEX_MODEL > config.toml
precedence; run-with-timeout supervises with the existing gtimeout/timeout/
watchdog chain and returns 124 on a deadline; an unknown subcommand exits 64.

Sourcing keeps working for skills rendered before the upgrade and prints one
deprecation line naming /gstack-upgrade; it and the zsh self-locate block are
removed in a release on or after 2026-10-21 (DX-18). Parity caps raised with
measured notes for the longer capture lines.

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* feat(gstack-post): one argv-only, redaction-scanned write path for PR and issue text

bin/gstack-post posts a PR/MR or issue comment, a review-comment reply, a
PR title or body, a new PR/MR or a new issue from agent-written files. Every
value reaches gh/glab as its own argv element in --flag=<value> form or on
stdin, never through a shell string, and a target is a number or a URL of
this repository on the same host.

The helper owns the redaction scan: it reads each file once, scans those
exact bytes (visibility from config, then gh/glab; the pusher's own email
suppressed) and sends the same bytes. HIGH is refused with exit 1 and no
confirmation can post it. MEDIUM exits 2 with RULE:/LINE:/PART: lines and a
TOKEN: digest of the bytes, destination, operation and findings; only
--confirm <token> for those exact bytes posts, so any edit needs a new
confirmation. gh/glab failures exit 3, usage errors 64. It also carries the
projectCards REST fallback and the title read-back that /ship did by hand.

The logic lives in lib/gstack-post.ts behind an injectable runner, so its
tests run in the Windows free lane; the stub-gh executable test is POSIX.

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* fix(autoplan): guard verifies dispatches on Claude Code 2.1.29x (schema strip, unflushed call, named causes)

Cause A: Claude Code journals the model's raw Agent input but hands
PreToolUse the schema-parsed input; with the fork-subagent gate on (the
interactive default in 2.1.292) run_in_background is dropped, so the journal
record never deep-equaled the payload. nativeToolInput now drops only that
documented strip for Agent. Agent input is bound on both paths: keys limited
to prompt, description, subagent_type (general-purpose) and
run_in_background; a model override or any other key is denied.

Cause B: when a ready journal lacks the current call, the guard evaluates
the payload placed after the last journaled record instead of polling for a
record that may arrive only with the tool result. A phase transition counts
a report only when a later journaled record follows it (and, on the journal
path, it is not in the current call's message). One batch rule covers both
paths; a background reviewer keeps the invocation armed until its
completion notice.

Every outcome now has a stable reason code, disposition and anchor
(autoplan/bin/guard-reasons.ts). UC1: Claude-Code-caused states
(unrecognized shape, journal lag, rewritten journal, oversized record,
unchecked-version schema strip) allow with a visible warning; integrity
failures stay denials. No denial advises a bare retry. Every guarded
decision appends one content-free line to analytics/autoplan-guard.jsonl,
trimmed to 1,000 lines past 256 KiB.

Co-authored-by: Jared <182125291+yolo-jared@users.noreply.github.com>
Co-authored-by: crblabs <293123359+crblabs@users.noreply.github.com>
Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* fix(autoplan): publish each phase report with a no-op call; foreground requests are conditional

Phase-close step 6 sends the report in a message whose only tool call is
the Bash no-op ': autoplan-published <phase>', and step 7 makes the next
guarded call in a later message, so the guard sees the report flushed
before the next phase entry regardless of Claude Code's journal timing.

Every site that asked for run_in_background: false now asks for it only
when the Agent schema offers it and says how to recover when a dispatch
runs in the background anyway (wait for its completion notice). Parity
caps rise by the measured bytes.

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* feat(ship,spec,review): PR and issue text goes out only through gstack-post

/ship: Step 18 writes the title into an agent-written file. An existing PR's
title is piped from gh/glab through gstack-pr-title-rewrite.sh --stdin (new),
so neither the agent nor a shell string carries it; a new title is written
with the file tool and its version prefix checked in the file. Step 19
composes the body as before and publishes with gstack-post pr-create, or
pr-body then pr-title, which scan the exact bytes they send. The separate
title and body scans, the hand-rolled REST fallback and the title self-check
are gone (the helper does all three). Title preparation moved into the
pr-body section, so the skeleton shrinks.

/spec files its issue with gstack-post issue-create and drops the pre-issue
re-scan; Greptile replies use gstack-post reply and pr-comment.
scripts/resolvers/redact-doc.ts loses its PR and issue sinks, and a test
keeps any template from rendering one again.

Tests that pinned the old scan-then-gh wiring now pin the gstack-post wiring;
the publication tests run the real compose and publish blocks against a stub
gh with the real helper (HIGH in title or body refused, MEDIUM token flow).

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* ci: pin Claude Code 2.1.292 in the CI image

2.1.292 is the npm latest users run. Its interactive default turns on the
fork-subagent gate, which strips run_in_background from the Agent schema
and backgrounds reviewers; CI on 2.1.284 never saw the /autoplan guard
break. The image tag is a content hash, so every eval workflow picks up
the new image without a workflow edit.

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* fix(brain-sync): one drain lock for every writer to the artifacts repo

gstack-artifacts-init, gstack-brain-restore, gstack-gbrain-source-wireup
and the skill-start merge ran git in $GSTACK_HOME without the drain's lock,
so the drain could never tell an abandoned .git/index.lock from a live one.
They now take the same mkdir lock (bin/gstack-brain-lock.sh): the drain and
the skill-start merge skip when it is held, user-run writers wait for it
(GSTACK_BRAIN_LOCK_WAIT, default 120 s), and a child of the holder (the
drain's daily worktree advance, restore's wireup) inherits it by token.

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* fix(brain-sync): hold back only the flagged file, keep coupled artifacts whole

One file with secret-shaped text used to block every push, silently. The
drain now attributes scanner hits to files in one pass over the staged diff
(split on its diff --git headers), holds back each flagged file with the
rest of its coupled group, unstages it, and requires the remaining staged
diff to re-scan clean; anything short of that falls back to a full unstage
and status blocked. Held files stay queued and are re-scanned every drain.

- Status file gains held_count, drainable, last_drain_at, last_push_at and
  held[] (path, scanner rule, dependents, both fixes); new status held.
- Coupled groups come from an explicit map (decisions.jsonl and
  decisions.active.json): a queued member stages the whole group, a group
  stages only when its snapshot matches its log in the index, a skipped
  member stops the whole group.
- --unskip-file undoes --skip-file and queues the path again.
- An index lock older than 10 minutes is removed under the drain lock; a
  younger one, a failed git add, a failed commit (hooks included) keep the
  queue and report status error. A verified empty staged diff is the only
  "nothing to commit". git add runs once, batched.
- The scanner reads bytes, so undecodable content no longer crashes it into
  a false clean; matched text never reaches the status file.
- An empty-queue run with unpushed sync commits stays push_failed.
- Privacy-held duplicate records collapse to one per path; --status counts
  a spool too deep for a shell glob.
- --attention prints skill start's fixed attention lines from the sync
  status table in the usage header.

Co-authored-by: v639dragoon <matthewsbrickley@gmail.com>
Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* feat(skill-start): ARTIFACTS_SYNC attention lines for a stuck sync

Skill start now prints one fixed attention line per sync-status-table row
that applies (held, blocked, push_failed, error, unknown, stale-push,
stale-drain), each naming gstack-brain-sync by absolute path. Values from
the status file are clamped (status code to the table, counts to digits,
timestamps to their shape), so the status message, file paths and scanner
output never reach the line. The preamble tells the model to surface the
line in one sentence and continue. Skill-start tests are generated from the
table in bin/gstack-brain-sync, and docs/gbrain-sync-errors.md carries the
same table.

Co-authored-by: v639dragoon <matthewsbrickley@gmail.com>
Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* fix(outside-review): severity words in label position count as findings

The Oct 7 /autoplan run's Codex CEO and DX reviews labeled seven and eight
findings High/Medium without [Pn] tags, and the classifier printed
VERDICT: clean / FINDINGS: none for both. lib/outside-review-result.ts now
reads severity words in label position (Severity: High, a leading High: /
High — / [High], **High**, a table cell): critical/high block like P0/P1,
medium/low are advisory. Words in prose (high-level, low-risk, no critical
findings) never count. A review or structured response with neither tags,
labels nor an explicit no-findings statement is unverified, never clean
(DX-11, ENG-10). Design-direction proposals move to a new proposal gate so
they keep needing only their Recommendation line, and review prompts ask
for a severity per finding. Fixtures reproduce the CEO/DX label shapes, the
eng [P1] output, a table and prose negative controls; the stored Codex panel
keeps its verdicts. docs/troubleshooting.md explains how the verdict is read.

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* feat: gstack-doctor readiness report and setup --status Codex row (B5)

bin/gstack-doctor prints one row per check without starting a skill:
install root (the registry rows behind ./setup --status), state root, Bun
against the floor, hook parse check, Codex mode and self-locate via
gstack-codex-probe subcommands, the cached model probe with its age
(--live runs the paid probe-model), artifacts-sync health from the status
file with the status code clamped, the browse bundle, Claude Code version,
the largest session journal and the last five schema-1 guard reason codes.
Rows are ok, warn, not configured or fail; only fail exits non-zero; fix
lines name helpers by absolute install-root path.

./setup --status gains a read-only Codex row (bin/gstack-codex-status.sh,
shared with the doctor) and ends with the doctor's absolute path.

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* docs: bug-report template asks for doctor output; README troubleshooting entry; doctor test in the Windows lane

The template gives the Claude Code doctor path and a lookup for other
hosts and names no bare gstack- command (asserted in
test/gstack-doctor.test.ts). The doctor test runs every script through
explicit bash argv, so it joins the curated Windows free subset.

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* docs(troubleshooting): readiness check section points at gstack-doctor

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* fix(question-tuning,plan-tune,gbrain): free text reaches helpers only through agent-written files

Question tuning no longer pipes the question summary through printf: a
registered id needs no summary, and for an unregistered id the agent writes
the summary to .gstack/tmp/qt.txt and passes --summary-file (new). The
helper resolves a relative path against the git top level, removes scratch
files under .gstack/tmp, and asks normally when the file is missing.

/plan-tune tune events carry the user's own words again: they go into a
file the agent writes and reach gstack-question-preference --write through
--free-text-file (new), still behind the injection filter and the
user-origin gate.

docs/gbrain-write-surfaces.md drops its heredoc templates for the
agent-written page file on stdin, matching the per-skill save block, and the
generic save block stops showing a quoted --content argument.

The tuning prose was tightened so every preamble stays within its size cap.

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* test(evals): daily autoplan-schema-canary on the latest Claude Code

A periodic rule case installs the latest published Claude Code with
install scripts off, verifies registry signatures, and runs one headless
session with the fork-subagent gate on whose PreToolUse hook records an
Agent and a Read payload. Each payload must match its journal record under
the guard's own comparison, keep the journal's key set apart from the
documented run_in_background strip, and stay inside the Agent allowlist;
a background Agent's completion notice must keep the shape the guard reads.
evals-periodic.yml runs it daily on its own schedule and concurrency group
(read-only token); the weekly census also includes it. A free test proves
the comparison fails on an injected strip.

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* fix(doctor): read the artifacts sync held count from held_count, the field the drain writes

* fix(artifacts sync): the init and restore pre-commit hooks catch Bearer/Basic/Token-prefixed JSON auth headers like the drain does

* feat(greptile-early): detect Greptile, ask once on public repos, open the early draft, wait for the review

bin/gstack-greptile-early backs /ship's early Greptile PR:
- detect: a root .greptile/ folder (which takes precedence), greptile.json,
  or a Greptile comment on the repo in the last 90 days; triggerOnDrafts from
  the effective root config; repo visibility; public-repo consent. It prints
  the signal, that /ship will push and open a draft now, the 10-minute wait
  cap and the absolute gstack-config command that turns it off.
- consent yes|no: the public-repo answer, remembered per repo in the
  project's state.
- open: opens the draft through gstack-post with a body saying /ship opened
  it early, and posts one @greptileai comment through gstack-post when the
  config does not set triggerOnDrafts: true. An open PR is left alone.
- wait: polls every 30 s for a completed Greptile check run, a Greptile
  review of the head commit or a summary naming it; each call returns within
  about 90 s and reports timeout 10 minutes after the PR opened.

gstack-config gains ship_greptile_early (default true, invalid values
rejected): on automatically whenever the repo uses Greptile, off per user.
The logic is in lib/greptile-early.ts behind an injectable runner, so its
fixture-repo tests run in the Windows lane; an end-to-end test drives the
real helper and gstack-post against a stub gh.

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* feat(ship): early Greptile PR at Step 6.5, bounded wait at Step 10 (#3020)

When the repo uses Greptile, /ship pushes and opens a draft PR right after
its free tests pass (new lazy section greptile-early, Step 6.5), so Greptile
reviews while Steps 7-9 run. A public repo asks once, remembered per repo;
declining keeps today's order. Step 10 waits for that review with
gstack-greptile-early wait and takes the existing unavailable route on a
timeout, never a claim of zero comments. Step 18 replaces the provisional
title, Step 19 marks the draft ready unless a draft was requested, and a
/ship that stops after Step 6.5 leaves the draft open with one comment
saying why.

ship's union cap rises to 1.5275 with the measured note (B4 +836 bytes,
B2 +4,732 bytes); the step-numbering guard allows 6.5; ship goldens are
refreshed for B4 and B2.

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* fix(autoplan): the publication no-op is 'true autoplan-published <phase>'

Measured on Claude Code 2.1.292 in a default-settings interactive session:
a bare ':' Bash command asks for approval ('This command requires
approval'), and after a background reviewer's completion notice starts a
new turn the skill's allowed tools no longer cover it, so the phase-close
no-op prompted the user. 'true' with arguments runs without approval and
prints nothing. The guard only needs a later journaled record, so the
command changes everywhere it is named.

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* test(evals): live autoplan-guard-pty case on the pinned Claude Code

A gate-tier rule case (one trial, haiku) drives an interactive PTY
session with default settings through a project skill that carries
/autoplan's own hook block: init, Phase 1 entry, a CEO reviewer dispatch
that sends run_in_background, a prepared close packet, the report
published with the no-op, and the Phase 2 entry in a later message. It
reads outcomes from the guard decision log (three verified allows) and
fails if the no-op raises a permission card; other cards are approved and
printed.

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* fix(autoplan): bounded, append-tolerant owned-journal read (A2)

The publication guard no longer reads the whole parent journal or refuses
it at 32 MiB. lib/claude-owned-journal.ts streams the bytes present at open
to the last complete line (pass one), keeps a content-free index per record
and decides ownership, ancestry and causal order on it with the same checks
and reason codes as before. Pass two reloads in full only what the
evaluation reads (autoplan/bin/guard-journal.ts): every literal init and its
result, the current call, and inside the invocation window (from the latest
human turn at or before the bound chain's first init) guarded and
close-packet Read results, structured Bash results, error results, assistant
text after the first close Read and active-plan mutations, plus the earlier
Read a dedup reply cites even before the window (ENG-4).

- lib/claude-journal-records.ts: per-record event extraction, record
  metadata and the causal-order checks, shared by readPlanCountTranscript
  and the bounded reader. The batch reader keeps its whole-file read and cap.
- Appends after open never fail a read; same device and inode and no
  shrink are required (identity). A second read in one hook invocation must
  hash bytes [0, first size) the same, or it is `rewritten` (UC1 unverified);
  the unrecognized-shape double read uses that prefix hash. `changing` is
  removed (REMOVED_CODES; its anchor stays).
- OWNED_RECORD_MAX_BYTES (32 MiB): only one record over it is too_large.
  OWNED_RETAINED_MAX_BYTES (32 MiB): over it is the new oversized_invocation
  fallback denial. GSTACK_TRANSCRIPT_TEST_MAX_BYTES and
  GSTACK_TRANSCRIPT_TEST_RETAINED_BYTES can only lower them.
- ENG-5: evaluation groups uses and results by tool id once (no per-tool
  rescans).

CEO-25 / ENG-5 spike (120 MiB synthetic, 31,242 records, 4 vCPU Linux,
Bun 1.4.2): a full streaming rescan (parse every line, small index, sha256)
took 254-326 ms at 147-161 MiB peak RSS; a pass-one index cache under the
state root (load cached index, verify a 64 KiB tail hash, parse the tail)
took 26-29 ms at 47 MiB. The rescan meets the 1 s budget, so the cache is
not adopted: it would add trusted persistent state and invalidation for no
budget need. A backward tail read cannot prove competing_root or cycle, so
it was not a candidate.

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* test(autoplan): journal-read benchmark and its dispatch-only workflow (CEO-2, CEO-6)

scripts/measure-journal-read.ts builds two synthetic 120 MiB parent journals
around a real /autoplan invocation (mixed: ~28k records, most bytes inside
the invocation; large-invocation: ~89k records with a current invocation of
many small records), keeps appending while the guard reads, and measures one
hook call in a fresh process (wall ms, peak RSS). The workflow runs it on
ubuntu-latest and macos-latest on workflow_dispatch only.

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* test(evals): autoplan-long-session, a resumed >100 MiB session enters Phase 1

CEO-15 with ENG-6's observer: a real session is compacted, its journal
padded past 100 MiB with owned records before the compact boundary, and
the resumed interactive session runs the scripted phase boundary; the
guard decision log must show a verified Phase 1 allow. A free test proves
that observer on a padded journal through the lower-only bound seams
(the PTY counting reader refuses the same journal), with a sidechain
control that must deny.

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* v1.91.34.0: CHANGELOG, TODOS, version; harness decision for test-pr-profile; ship title-step test reads its new section

* docs(changelog): use an example.com address in the allowEmail example

* test: assemble GitHub-token fixtures at runtime in the new brain-sync and skill-start tests (CI credential scan)

* fix: integration repairs across the wave (Step 6.5 fence, printed helper paths on Windows, probe usage without a heredoc) and tests pinned to the moved title and gstack-post contracts

* test(doctor): the healthy fixture builds browse.exe on Windows, the name setup uses there

* test(shared-libs): recognize Claude Code 2.1.292's internal git probe prefix (format.pretty and per-hook disables), with its CI receipt

* fix(ship-measure): report lists v1 measurements as an older format instead of crashing

---------

Co-authored-by: Mike Ilog <48659643+Mike-E-Log@users.noreply.github.com>
Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>
Co-authored-by: Thomas Gernhardt <t.gernhardt@softwareprojekt.net>
Co-authored-by: akmandhania <143462686+akmandhania@users.noreply.github.com>
Co-authored-by: Jared <182125291+yolo-jared@users.noreply.github.com>
Co-authored-by: crblabs <293123359+crblabs@users.noreply.github.com>
Co-authored-by: v639dragoon <matthewsbrickley@gmail.com>
2026-10-07 23:29:05 +00:00

666 lines
33 KiB
Cheetah
Raw Permalink Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
---
name: ship
preamble-tier: 4
version: 1.0.0
description: |
Ship workflow: detect + merge base branch, run tests, review diff, bump VERSION,
update CHANGELOG, commit, push, create PR. Use when asked to "ship", "deploy",
"push to main", "create a PR", "merge and push", or "get it deployed".
Proactively invoke this skill (do NOT push/PR directly) when the user says code
is ready, asks about deploying, wants to push code up, or asks to create a PR. (gstack)
allowed-tools:
- Bash
- Read
- Write
- Edit
- Grep
- Glob
- Agent
- AskUserQuestion
- WebSearch
sensitive: true
triggers:
- ship it
- create a pr
- push to main
- deploy this
---
{{PREAMBLE}}
{{THIRD_PARTY_ACTIONS}}
# Ship: Fully Automated Ship Workflow
STOP blocks advancement until the stated repair/resume route clears; without one, end this attempt.
Answer each AskUserQuestion before continuing.
Routine authorization never waives those gates or their required user decisions.
**Routine work needs no confirmation:** include uncommitted changes, choose MICRO/PATCH
under Step 12, draft CHANGELOG and commits, mark completed TODOs and auto-fix findings.
When Step 7 coverage meets its target, report remaining gaps and verify generated
tests without another permission question. Step 15 commits those tests.
**Route:** integrate (1–3) → test and review (4–11.5) → prepare the release
(12–15) → verify frozen content (16) → push and publish (17–21).
Every new invocation repeats Steps 1–16, including both reviews and the docs audit.
Steps 12, 17 and 19 prevent duplicate bumps, pushes and PRs, never verification.
### Keep state between steps
Keep one private Markdown **invocation record** outside the product tree and save
its absolute path. Use these headings so a paused run can resume:
- **Release:** versions, `BUMP_LEVEL`, reviewed tree and attempt counts.
- **Decisions:** each approval's finding, files and authorized action. Reuse it only
for that same scope; a repair never resets approvals or expands them.
- **Reviews:** handles, original start tokens, terminal states, outputs and queued fixes.
- **Checks:** command/label, result/counts, timestamp, log and consumed inputs.
- **Documentation:** candidate/id, attempts used, accepted hashes or named blocked exception.
- **Next steps:** one ordered work list, with the current step marked.
A **receipt** is saved evidence of a check's command, result and consumed content.
A review's **start token** is the opaque value returned by `gstack-review-log --start`
before it reads the diff. Keep `REVIEW_START` for Step 9, a separate `PASS_START` for
each Step 11 attempt, and `DESIGN_START` for design. Finish each pass with its original
token; `--finish` stamps the binding fields automatically. Never borrow or replace a token.
`gstack-wtree` prints a Git tree hash covering tracked and non-ignored untracked files,
not a commit ID. Use `git diff <old-tree> <new-tree>` to compare these snapshots.
### Ship control flow
You, the **parent** running /ship, own advancement; children return evidence, not
permission to proceed. Follow the saved work list:
1. Start with Steps 1–21 in order, including 6.5, 11.5 and 14.5. Advance only after
the current item's gates clear.
2. Expand a repair into individual steps and insert them before the still-pending
work. This replaces the current item, whose actual result stays in the record.
Add its destination only if not already the next pending step.
3. For another repair, repeat rule 2 without discarding pending work.
The saved list takes precedence over ordinary next-step
sentences inside a repair. A range never adds unlisted steps.
**Example:** Step 11 fixes insert `9 → 10 → 11` before 11.5. A further Step 9 fix
affecting 6–8 makes the list `5 → 6 → 7 → 8 → 9 → 10 → 11 → 11.5`.
The unchanged release steps follow. STOP and AskUserQuestion gates still apply during repairs.
Keep the same attempt counts throughout the invocation. A range ending at Step 14
does not enter Step 14.5. A range that includes Step 14.5 enters its existing audit
decision, not an unconditional new launch; its initial-plus-ONE limit never resets.
Permitted repairs continue in this invocation without restarting /ship.
---
{{SECTION_INDEX:ship}}
---
{{BASE_BRANCH_DETECT}}
`<base>` means the detected branch name for fetch/helper arguments;
`origin/<base>` is its remote-tracking ref for comparisons. Step 1 fetches it.
{{GBRAIN_CONTEXT_LOAD}}
## Step 0.9: Apple target detection
If the ask is App Store/TestFlight distribution, look for an `.xcodeproj`,
`.xcworkspace`, or Swift app product. Read `Package.swift` and its entrypoint to
distinguish an app from a library/CLI. If unclear, use AskUserQuestion to identify
the target and wait before choosing a release path.
For a confirmed app, **STOP and Read
`~/.claude/skills/gstack/ship/sections/apple-release.md` FIRST**. Store distribution proceeds
through that adapter from the current branch, including a clean base branch.
The branch gate and repository-landing pipeline below apply ONLY to
repository-landing asks, including on Apple repos.
## Step 1: Pre-flight
1. Save the current branch as `<branch-name>`. If on the base branch or the repo's default branch, **abort**: "You're on the base branch. Ship from a feature branch."
2. Run `git status` (never use `-uall`). Uncommitted changes are always included — no need to ask.
3. Run `git fetch origin <base>` before inspecting the diff. If fetch fails, STOP:
report the error and restore access before continuing. Then inspect
`git diff origin/<base> --stat`, untracked files from status, and
`git log origin/<base>..HEAD --oneline`.
4. Display historical readiness using the dashboard below, then finish Step 1.
Prior CLEAR reviews or dashboard skips never replace Step 9's gates.
{{REVIEW_DASHBOARD}}
An outside review with status `unverified` or `unavailable` is missing coverage, never a pass.
For diffs >200 lines (`git diff origin/<base> --stat | tail -1`), recommend
`/plan-eng-review` or `/autoplan` for architecture review.
For Design Review: run `source <(~/.claude/skills/gstack/bin/gstack-diff-scope <base> 2>/dev/null)`. If `SCOPE_FRONTEND=true` and no design review exists, mention: "Design Review not run — Step 9 includes the lite check; consider /design-review for a full visual audit."
Continue to Step 2 without asking; Step 9 applies the review gates.
---
## Step 2: Distribution Pipeline Check
Check distribution for new standalone artifacts (CLI binaries, packages, tools),
not web services with existing deployment.
1. List candidate distribution paths:
```bash
git diff origin/<base> --diff-filter=A --name-only | grep -E '(^|/)(cmd/[^/]+/main\.go|bin/[^/]+|Cargo\.toml|setup\.py|package\.json)$' | head -5
```
Also inspect matching untracked files from Step 1's status. Read each match:
a new `package.json` or `Cargo.toml` alone does not establish a publishable
artifact. Also inspect existing manifests for newly declared binaries or
package exports. Apply the pipeline gate only when a new distributable is present.
2. If new artifact detected, check for a release workflow:
```bash
ls .github/workflows/ 2>/dev/null | grep -iE 'release|publish|dist'
grep -qE 'release|publish|deploy' .gitlab-ci.yml 2>/dev/null && echo "GITLAB_CI_RELEASE"
```
3. **New artifact without a pipeline:** AskUserQuestion: "Users cannot download this
artifact after merge without a release pipeline."
- A) Add the platform's release workflow now
- B) Defer with a P1 distribution TODO in Step 14
- C) Not needed: internal/web-only, covered by existing deployment
4. **If A:** Add packaging/publish configuration using repository CI conventions.
Ask for unknown targets, registries or access first; never invent credentials.
Recheck against the artifact and include the workflow in tests and review.
Do not publish a release during `/ship`.
5. Otherwise, continue without adding a pipeline.
---
## Step 3: Merge the base branch (BEFORE tests)
Merge the base ref fetched in Step 1 so tests and reviews cover the integrated code:
```bash
git merge origin/<base> --no-edit
```
**If there are merge conflicts:** Try to auto-resolve if they are simple (VERSION, schema.rb, CHANGELOG ordering). For complex or ambiguous conflicts, **STOP**, show the conflicting choices, use AskUserQuestion for the needed resolution decision, and wait for the answer before editing or continuing.
**If already up to date:** Continue silently.
If integration changes the artifact or distribution configuration inspected in Step 2,
repeat Step 2 on the merged content, including its decisions, then continue to Step 4.
Otherwise continue to Step 4 directly.
---
{{SECTION:tests}}
{{SECTION:greptile-early}}
{{SECTION:test-coverage}}
{{SECTION:plan-completion}}
{{SECTION:review-army}}
{{SECTION:greptile}}
{{SECTION:adversarial}}
## Step 11.5: Bind the reviews
1. **Select the two reviews.** Run `~/.claude/skills/gstack/bin/gstack-review-read`.
Select this invocation's final Step 9.4 record (`skill:"review"`, `via:"ship"`)
and Step 11 native record (`skill:"adversarial-review"`). Match each to its saved
handle, original token and source; reject outside-provider or older invocation records.
2. **Compare their content.** Require the native record's `review_binding.state`
to be `verified`. All three snapshots must match: its `wtree`, Step 9.4's
`review_binding.start_wtree` and `review_binding.end_wtree`. A mismatch or missing
record/field blocks release preparation: report **Review records missing or mismatched**
and insert `9 → 10 → 11 → 11.5` before Step 12. Bind the new records at 11.5.
Never attach new tokens to old work.
3. **Preserve any QA exception.** A named probe-risk exception may leave Step 9.4's
root `wtree` absent; item 2 still compares its start/end snapshots. Matching content
does not mean the failed or unrun probes passed. Keep Step 9.4's incomplete flags
and the user's exception.
4. **Save the evidence.** Save both records and matching **reviewed tree** for
Step 16. Continue to Step 12.
## Step 12: Version bump (auto-decide)
Item 3 needs `BUMP_LEVEL`: reuse this invocation's saved level. Otherwise FRESH
chooses it in item 2 and ALREADY_BUMPED derives it in item 1.
1. **Classify state** — pure reader, never writes:
```bash
bun run ~/.claude/skills/gstack/bin/gstack-version-bump classify --base <base>
```
Save the JSON `baseVersion` as `BASE_VERSION`, then read `state` and dispatch:
- **FRESH** → use the recorded level or choose it in item 2, then check the queue and write.
- **ALREADY_BUMPED** → keep `NEW_VERSION=currentVersion`. If `BUMP_LEVEL` is missing,
use the first changed component from `baseVersion` to `currentVersion`
(major/minor/patch/micro; an absent fourth component is zero). Continue at item 3,
not another automatic bump.
- **DRIFT_STALE_PKG** → run `gstack-version-bump repair`, then reclassify.
Success follows ALREADY_BUMPED, including its queue check; failure stops.
Repair alone never re-bumps.
- **DRIFT_UNEXPECTED** → STOP: package.json disagrees with VERSION while VERSION
matches base. Reconcile the manual edit, then reclassify.
- **NO_VERSION** → print `notice` verbatim and ship without a version change: skip
the rest of Step 12 and Step 13's CHANGELOG entry, never create VERSION, use an
unprefixed title in Step 18 and log `"version":null` in Step 20.
- **Exit 2** → STOP and show stderr: a configured version file is missing, empty,
unreadable or malformed. Fix it or its pin, then reclassify. Never substitute `0.0.0.0`.
2. **Decide the bump level** from the diff (agent judgment):
- **MICRO**: <50 lines, trivial tweaks/config. **PATCH**: 50+ lines, no feature signals.
- **MINOR**: ask for any feature signal (new route/page, migration, module) or 500+ lines.
**MAJOR**: ask for milestones or breaking changes. Use AskUserQuestion: recommended
level with rationale, smaller level, or cancel. Wait; cancel stops before release
writes or push and preserves existing work.
Save lowercase `BUMP_LEVEL`. A claimed version may move the next available number
forward, but cannot change the chosen MICRO/PATCH/MINOR/MAJOR level.
3. **Queue-aware pick** (workspace-aware ship):
```bash
QUEUE_JSON=$(bun run ~/.claude/skills/gstack/bin/gstack-next-version --base <base> --bump "$BUMP_LEVEL" --current-version "$BASE_VERSION" 2>/dev/null || echo '{"offline":true}')
CANDIDATE_VERSION=$(echo "$QUEUE_JSON" | jq -r '.version // empty')
```
**Qualify first:** require successful utility output and a nonempty valid version.
`offline:false` qualifies; `offline:true` qualifies only with `fallback:"git"`.
Offline output without that fallback, failure, malformed output or an empty version
is unusable, even if it contains a version-looking string.
- **Usable candidate:** print warnings and claimed queue. FRESH sets `NEW_VERSION=CANDIDATE_VERSION`.
ALREADY_BUMPED compares it with `currentVersion`: if different, ask to rebump
(refresh CHANGELOG/PR title) or keep current (CI rejects a collision).
Only approval changes the existing version. Check JSON `active_siblings` by
`branch` and `version`; a sibling holding `>= NEW_VERSION` requires a choice:
advance past it, or stop this attempt and sync.
- **No usable candidate:** print queue-unverified. FRESH uses local `BUMP_LEVEL`
arithmetic; ALREADY_BUMPED keeps `currentVersion`. Never use an empty candidate.
4. **Write the bump** (FRESH, or an approved rebump):
```bash
bun run ~/.claude/skills/gstack/bin/gstack-version-bump write --version "$NEW_VERSION" --regen-digest
```
The CLI validates `MAJOR.MINOR.PATCH.MICRO` (or pinned 3-digit semver) and writes
VERSION, the manifest and existing `package-lock.json` / `npm-shrinkwrap.json`;
it never creates lockfiles. Manifest path: `--package-json-path` →
`.gstack/package-json-path` → `./package.json`. npm files use the 3-digit translation
(`1.67.0.0` → `1.67.0`); VERSION is authoritative. Exit 3 means a half-write:
reclassify and `repair` DRIFT_STALE_PKG.
`--regen-digest` runs repo code with Step 5's privileges: `scripts/gen-agents-digest.ts`,
only when it and committed `agents-digest/gstack-AGENTS.md` exist. If `agentsDigest`
is false, run `bun scripts/gen-agents-digest.ts` and stage the digest with the bump.
Before push, verify the committed digest matches generation for the selected VERSION.
5. **Record the release decision after a version was actually written**, including
an approved ALREADY_BUMPED rebump. Skip unchanged versions and manifest-only repairs.
```bash
~/.claude/skills/gstack/bin/gstack-decision-log '{"decision":"Ship NEW_VERSION (BUMP_LEVEL)","rationale":"WHY","scope":"repo","source":"skill","confidence":9}' 2>/dev/null || true
```
Substitute `NEW_VERSION`, `BUMP_LEVEL`, and one-line `WHY` (scope or breaking-change signal). Best-effort, non-interactive, non-blocking.
{{SECTION:changelog}}
## Step 14: TODOS.md (auto-update)
Read `~/.claude/skills/gstack/review/TODOS-format.md`.
**1. Open or create:** Read root `TODOS.md`. An explicit "add TODO" choice authorizes
creation with `# TODOS` and `## Completed`. Otherwise, if missing, ask: A) Create
a component/priority-organized TODOS.md, B) Skip. Skip goes to item 5.
**2. Organization:** Use component headings, `**Priority:**` P0–P4 and `## Completed`
at the bottom. If disorganized, ask: A) Reorganize preserving all content
(recommended), B) Leave as-is.
**3. Add approved deferrals:**
- Step 2: add the approved distribution follow-up as P1 with the missing pipeline and affected artifact.
- Step 8: add each approved P1 plan deferral with `Deferred from plan: {plan file path}` and the missing work.
- Step 5: retain P0 test-failure entries already written; deduplicate by failure and source, adding missing approved entries with error output and branch.
Never turn dropped scope into TODOs or invent unapproved follow-ups. Reuse matching existing entries rather than duplicating them.
**4. Detect completed TODOs:** Compare titles, files and behavior with
`git diff origin/<base>`, untracked files and `git log origin/<base>..HEAD --oneline`.
Move proven completions to `## Completed` with `**Completed:** vX.Y.Z (YYYY-MM-DD)`;
leave uncertain items open.
**5. Save the summary:** Report additions, deferrals, completions, remaining count and
creation/reorganization. If creation was declined or a write failed, warn and retain
unsaved follow-ups in Step 19's PR summary. Never claim they were saved;
TODO write failures are non-blocking.
---
## Step 14.5: Documentation audit (every ship)
**Doc-sync invariant:** Every ship dispatches the /document-release subagent before final
commit/verification/publication, including reruns, already-pushed branches, existing PRs and docs-only changes.
No edits means an executed audit, not a skip; report the section's verified outcome.
{{SECTION:documentation}}
## Step 15: Commit (bisectable chunks)
Make bisectable commits; if already committed, continue to Step 16. Never create an empty commit.
1. Group changes with their tests, config/routes, views and Step 14.5 docs.
Migrations may stand alone or accompany their model.
Under 50 lines across fewer than 4 files may use one commit.
2. Order dependencies first: infrastructure → models/services → controllers/views.
Each commit must work independently, without broken imports or missing code.
Group VERSION + CHANGELOG + TODOS.md after the feature commits.
3. Use `<type>: <summary>` (feat/fix/chore/refactor/docs) and a brief body.
Only the final VERSION/CHANGELOG commit gets the release version and co-author
trailer. Do not create a Git tag:
```bash
git commit -m "$(cat <<'EOF'
chore: bump version and changelog (vX.Y.Z.W)
{{CO_AUTHOR_TRAILER}}
EOF
)"
```
---
## Step 16: Verification Gate
**IRON LAW: NO COMPLETION CLAIMS WITHOUT FRESH VERIFICATION EVIDENCE.**
Run stages 1–5 in order. Recovery instructions below name where to resume.
If content changes during or after verification, restart at stage 1 and complete
all five stages before Step 17. Content-preserving commits keep valid evidence.
### 1. Finish writers and prepare outputs
Inspect writer handles, including the docs child. Confirm terminal completion or termination
before another writer runs. Timeout or cancellation acknowledgment alone means
STOP until confirmed.
Find declared generation/build commands in project instructions, manifests, build
files and CI. Run them and save results. If none exists, record not applicable and
the inspected sources. A missing prerequisite or failed build stops shipping:
report **Build failed or prerequisite missing**, with the command, error and needed
repair. Never invent a substitute command.
**If blocked:** Repair the prerequisite or build, then repeat stage 1. After it passes, continue
to stage 2; treat any content repair as a behavioral change there.
### 2. Choose the change route
Capture the current tree with `~/.claude/skills/gstack/bin/gstack-wtree`. Inspect
`git diff <reviewed-tree> <current-tree>` against the snapshot saved before Step 12.
Missing snapshots block this comparison, regardless of HEAD equality.
Classify the comparison in this order:
1. **Behavior, tests or build inputs changed:** Prompts/templates count as behavior.
Insert `5–11.5 → 12–14 → 16` before the pending Step 17, then stop this step.
This repair excludes Step 14.5 because the rebuild can change generated docs.
Step 16 restarts at stage 1: rebuild and compare again before stage 3 decides
documentation freshness. Further repairs use the same work list.
2. **Only authored docs or release metadata changed:** Keep Step 8's original child
report and counts. Recheck affected plan items using their recorded verification
and append current evidence to the invocation record. If a classification is no
longer supported, run Step 8's audit and decision gates only, then return to
Step 16 stage 1. Never edit the child's counts yourself.
3. **No changes, or the docs-only checks still support the plan:** Continue to stage 3
without a new code review.
### 3. Resolve documentation freshness
Compare the base and hashes of the selected release paths, generated
outputs and docs/templates with Step 14.5's saved values. A prior invocation's
audit or risk decision never qualifies.
| Outcome | Action |
|---|---|
| This invocation's accepted audit matches all inputs | Continue to stage 4. |
| User-accepted named documentation risk covers the same approved scope and exact content, and unwaivable gates clear | Continue to stage 4; retain `Documentation: blocked`, its reason and incomplete scope. |
| Missing, stale or blocked | Use recovery below. Never silently refresh hashes. |
Report changed inputs, blockers and attempts used:
- **An attempt remains, with changed inputs or an available repair:** insert
`14.5 → 15 → 16` before Step 17. Use Blocked recovery with the existing count.
Validate the outcome before Step 15,
then restart Step 16 stage 1 to regenerate and compare again.
- **Otherwise:** STOP unless the user accepts
the specific named documentation risk and all unwaivable gates clear, under
Step 14.5's Blocked recovery rules. Unchanged approved content goes to stage 4;
repaired content goes to stage 1.
Never run a third audit. Child return is not acceptance.
### 4. Verify the frozen candidate
Freeze inputs through verification and push. Run declared docs/link/generated-file
checks; report unavailable checks.
**Reuse a check when its inputs match.** Compare hashes or complete bytes of its
saved and current consumed files, fixtures, dependencies and execution parameters.
Explain why other changes cannot affect it; changed or unknown dependencies require a rerun.
For model judges, compare the complete expanded request, rubric, parameters and
builder/runtime dependencies. Reuse identical passing evidence: cite the original
command, result/counts, timestamp and log, never resample it. Mandatory reviews still run.
**Check each test lane's receipt as well.** Use its actual Step 5 label/command:
`--label <lane> --expect-cmd '<exact Step 5 command>'`. Inspect changes since the run;
`--allow-paths` exempts only release metadata. A `package.json` version-only edit
can qualify; scripts, dependencies and runtime configuration require live tests.
Uncertain edits cannot be exempted. Docs, TODO edits, new/generated tests and fixes
make evidence STALE even without a new code review. Use this example only after
confirming that every allowed edit is release metadata:
```bash
~/.claude/skills/gstack/bin/gstack-evidence check --label tests --expect-cmd '<tests>' --label vitest --expect-cmd '<vitest>' --max-age 24 --allow-paths CHANGELOG.md,VERSION,package.json,agents-digest/gstack-AGENTS.md
```
| Receipt result | Next action |
|---|---|
| FRESH (exit 0) | Cite the label, exit, timestamp and log. |
| STALE/MISSING: changed content, command or age, or no proven run | Run `~/.claude/skills/gstack/bin/gstack-evidence run --label <lane> -- '<command>'`, read the result and recheck once. Handle failures as described below. |
| Only receipt storage/readback failed | Independently prove unchanged final content, the same command and valid age from the successful run's evidence. Cite its exact command, exit, timestamp and log as **ledger unavailable**, never FRESH. Without that proof, use STALE/MISSING. |
No test lanes: require Step 5's explicit untested-scope approval for final content,
or run Steps 5–15, including the no-tests decision, then return to Step 16 stage 1.
Report the gap, never FRESH; builds must pass.
**New, changed or unwaived test failure:** STOP publication. Run Steps 5–15,
starting with Step 5's triage, then return to Step 16 stage 1. This recovery also
applies if a failure appears while reporting in stage 5. Reentry to Step 14.5
keeps its existing audit count; it does not authorize a third attempt.
### 5. Report, then push
Commit only approved, verified release changes left uncommitted after Step 15,
including generated outputs; use its grouping rules and never create an empty commit.
Preserve unrelated user files.
Paste build/docs/test results. Reuse waivers only for the same verified
pre-existing failures and approved scope; cite the actual approval and failing
counts, never FRESH or all-green. A new, changed or unwaived test failure uses
stage 4's recovery before publication. Otherwise continue to Step 17.
---
## Step 17: Push
**Credential pre-push guard — run before the push:**
```bash
_REDACT_PREPUSH=$(~/.claude/skills/gstack/bin/gstack-config get redact_prepush_hook 2>/dev/null || echo "false")
_HOOK_PATH=$(git rev-parse --git-path hooks/pre-push 2>/dev/null || echo "")
_HOOK_STATE="missing"
if [ -e "$_HOOK_PATH" ] || [ -L "$_HOOK_PATH" ]; then
_HOOK_STATE="unmanaged"
if [ -f "$_HOOK_PATH" ] && [ ! -L "$_HOOK_PATH" ] && grep -Fqx '# gstack-redact pre-push (managed)' "$_HOOK_PATH" 2>/dev/null; then
_HOOK_STATE="managed"
fi
fi
_HOOKS_DIR=$(git rev-parse --git-path hooks 2>/dev/null || echo "")
_HOOKS_IN_GIT_DIR="no"
_HOOKS_CONFIG_STATUS=0
git config --get core.hooksPath >/dev/null 2>&1 || _HOOKS_CONFIG_STATUS=$?
if [ -n "$_HOOK_PATH" ] && [ -n "$_HOOKS_DIR" ] && [ "$_HOOKS_CONFIG_STATUS" = "1" ] && [ ! -L "$_HOOKS_DIR" ]; then
_HOOKS_IN_GIT_DIR="yes"
fi
GSTACK_STATE_ROOT=$(~/.claude/skills/gstack/bin/gstack-paths --get GSTACK_STATE_ROOT); : "${GSTACK_STATE_ROOT:?gstack-paths failed; reinstall with ./setup or /gstack-upgrade}"
_PREPUSH_PROMPTED=$([ -f "$GSTACK_STATE_ROOT/.redact-prepush-prompted" ] && echo "yes" || echo "no")
if [ "$_REDACT_PREPUSH" = "true" ] && [ "$_HOOKS_IN_GIT_DIR" = "yes" ] && [ "$_HOOK_STATE" != "unmanaged" ]; then
~/.claude/skills/gstack/bin/gstack-redact install-prepush-hook || exit $?
fi
echo "REDACT_PREPUSH: $_REDACT_PREPUSH"
echo "HOOK_STATE: $_HOOK_STATE"
echo "HOOKS_IN_GIT_DIR: $_HOOKS_IN_GIT_DIR"
echo "PREPUSH_PROMPTED: $_PREPUSH_PROMPTED"
```
Branch on the echoed values:
1. **`REDACT_PREPUSH: true`** — the block installs or refreshes managed
hooks, preserving `pre-push.local` and complete stdin. On installer
failure, STOP before pushing. `HOOKS_IN_GIT_DIR: no`: do not install;
request manual integration. `HOOK_STATE: unmanaged`: ask consent only
for a regular, non-symlink hook in the default directory without
`pre-push.local`; otherwise request manual integration. Dangling
symlinks are unmanaged. Never overwrite either policy.
2. **`REDACT_PREPUSH` not true AND `PREPUSH_PROMPTED: no`** — one-time
offer (fires once EVER, machine-wide). AskUserQuestion:
> gstack can install a per-repo git pre-push hook that blocks pushes
> containing credentials (API keys, tokens, private keys). It's a
> guardrail, not enforcement — `GSTACK_REDACT_PREPUSH=skip` bypasses it.
> Install it for repos you ship from?
Options:
- A) Yes — install the credential guard (recommended)
- B) No — never ask again
If A: run `~/.claude/skills/gstack/bin/gstack-config set redact_prepush_hook true`
then re-run the block and apply the same directory and unmanaged-hook rules above.
If B: run `~/.claude/skills/gstack/bin/gstack-config set redact_prepush_hook false`.
ALWAYS (after either answer, but NOT if the question itself failed to
render — a failed AskUserQuestion must re-offer next time):
```bash
GSTACK_STATE_ROOT=$(~/.claude/skills/gstack/bin/gstack-paths --get GSTACK_STATE_ROOT); : "${GSTACK_STATE_ROOT:?gstack-paths failed; reinstall with ./setup or /gstack-upgrade}"
touch "$GSTACK_STATE_ROOT/.redact-prepush-prompted"
```
3. **Declined earlier** — continue
without comment.
**Idempotency check:** Check if the branch is already pushed and up to date.
```bash
LOCAL=$(git rev-parse HEAD) || exit 1
REMOTE_REF=$(git ls-remote --heads origin refs/heads/<branch-name>) || {
echo "STATUS: BLOCKED — cannot verify remote branch; restore access before pushing"
exit 1
}
REMOTE=$(printf '%s\n' "$REMOTE_REF" | awk '{print $(1)}')
REMOTE=${REMOTE:-none}
echo "LOCAL: $LOCAL REMOTE: $REMOTE"
[ "$LOCAL" = "$REMOTE" ] && echo "ALREADY_PUSHED" || echo "PUSH_NEEDED"
```
If `ALREADY_PUSHED`, skip the push but continue to Step 18. Otherwise push with upstream tracking:
```bash
git push -u origin <branch-name>
```
**If the push fails, STOP.** No Step 19 or publication claim. Report the error:
- **Non-fast-forward push:** fetch and inspect the remote, then merge under Step 3's
conflict rules. Run Steps 5–16 before returning to Step 17. Never rewrite history.
- **Authentication, hook or network failure:** repair the cause, then repeat Step 16
even if content is unchanged before returning to Step 17. Never bypass failed guards.
Never force-push.
Only a successful push or verified `ALREADY_PUSHED` proceeds.
Continue to Step 18. No documentation writer runs after push.
---
## Step 18: Prepare publication metadata
First look up open PRs/MRs for `<branch-name>` on the detected platform:
- GitHub: `gh pr list --head <branch-name> --state open --json number,title,url`
- GitLab: `glab mr list --source-branch <branch-name> --output json` (defaults to open).
A successful empty array means new; one match supplies the existing title/identity.
Lookup failure or ambiguous matches **STOP** for resolution, never mean no PR.
Save the result for Step 19's recheck.
The pr-body section prepares the title from that result, then publishes.
{{SECTION:pr-body}}
## Step 20: Persist ship metrics
Log metrics for `/retro` through `gstack-review-log`; it handles project/branch paths,
JSON validation, storage and sync. It takes **no path argument**; do not build one.
```bash
~/.claude/skills/gstack/bin/gstack-review-log '{"skill":"ship","timestamp":"'"$(date -u +%Y-%m-%dT%H:%M:%SZ)"'","coverage_pct":COVERAGE_PCT,"coverage_schema":2,"coverage_pct_value":COVERAGE_PCT_VALUE,"weak_gaps":WEAK_GAPS,"tests_extended":TESTS_EXTENDED,"tests_rejected":TESTS_REJECTED,"regression_proof":REGRESSION_PROOF,"plan_items_total":PLAN_TOTAL,"plan_items_done":PLAN_DONE,"verification_result":"VERIFY_RESULT","version":"VERSION","branch":"'"$(git rev-parse --abbrev-ref HEAD)"'"}'
```
Substitute from earlier steps:
- **COVERAGE_PCT**: Step 7 diagram's integer percentage; encode null/undetermined as -1
- **COVERAGE_PCT_VALUE**: Step 7's `coverage_pct_value` (the gate's X) as an integer, or `null` when missing or ignored
- **WEAK_GAPS**, **TESTS_EXTENDED**, **TESTS_REJECTED**: counts of Step 7's `weak_gaps`, `tests_extended` and `tests_rejected` (0 when the key is missing or ignored)
- **REGRESSION_PROOF**: `{"red_at_head":N,"base_green":N,"base_unavailable":N}` from Step 7, or `null` when missing
- **PLAN_TOTAL**: total plan items extracted in Step 8 (0 if no plan file)
- **PLAN_DONE**: count of DONE + CHANGED items from Step 8 (0 if no plan file)
- **VERIFY_RESULT**: "pass", "fail", or "skipped", set after Step 9 executes Step 8.1's verification list
- **VERSION**: `NEW_VERSION` from Step 12; `null` (unquoted) under NO_VERSION
The shell supplies the branch. Run this automatically, without confirmation.
---
## Step 21: Plan-tune discoverability nudge (first-successful-ship only)
After a successful ship, show the non-blocking /plan-tune nudge once per machine:
```bash
GSTACK_STATE_ROOT=$(~/.claude/skills/gstack/bin/gstack-paths --get GSTACK_STATE_ROOT); : "${GSTACK_STATE_ROOT:?gstack-paths failed; reinstall with ./setup or /gstack-upgrade}"
_NUDGE_MARKER="$GSTACK_STATE_ROOT/.plan-tune-nudge-shown"
_QT=$(~/.claude/skills/gstack/bin/gstack-config get question_tuning 2>/dev/null || echo "false")
if [ ! -f "$_NUDGE_MARKER" ] && [ "$_QT" = "false" ]; then
echo ""
echo "gstack can learn from your AskUserQuestion answers. Run /plan-tune to opt in"
echo "— it captures which prompts you find valuable vs noisy and (with hooks installed)"
echo "auto-decides your never-ask preferences."
mkdir -p "$GSTACK_STATE_ROOT" && touch "$_NUDGE_MARKER"
fi
```
The marker or enabled question_tuning suppresses it. To re-enable, remove
`$GSTACK_STATE_ROOT/.plan-tune-nudge-shown` before the next ship.
---
## Section self-check (before you finish)
List the applicable Section index entries and confirm each Read. If you worked from
memory, STOP, Read the section and redo that step. Use `gstack-version-bump`, never
hand-roll VERSION/package.json writes.
---
## Important Rules
Follow the numbered gates and their explicit exceptions.
- **Never force push.** Use regular `git push` only.
- **An early Greptile PR stays up.** A stop after Step 6.5 opened it leaves the draft
open with one comment saying why (greptile-early section); never close it.
- **Use the configured version file's format** (4-digit for VERSION); under NO_VERSION,
never invent one.
- **Step 7 generates coverage tests.** They must pass before committing. Never commit failing tests.