ci(infra): port four CI governance workflows (#37511)

Four GitHub Actions workflows ported from the Deep Agents monorepo to
enforce repository hygiene rules that were not previously applied here.

## Changes

- **Fork-main PR guard**: closes PRs from forks whose head is `main` or
`master`, with a sticky comment explaining how to reopen from a feature
branch. Prevents the "Update branch" → admin-override path that lets a
`Merge branch 'master' into master` commit land on the default branch
and bypass squash-only policy. Maintainers can override with a
`bypass-fork-main-check` label.
- **Monthly uv pin bump**: opens a PR on the first of each month to
advance `UV_VERSION` in the composite setup action. Probes
`releases.astral.sh` across four architectures before committing so CI
doesn't race a lagging mirror on fresh-release days — the gap
Dependabot's `github-actions` ecosystem can't cover because it tracks
`uses:` SHA pins, not the inline `UV_VERSION` value.
- **Extras-sync validation**: a Python script (`check_extras_sync.py`)
and companion workflow that detect version-constraint drift between
`[project.dependencies]` and `[project.optional-dependencies]` across
every `libs/**/pyproject.toml`. Runs on PRs touching any
`pyproject.toml` and on pushes to `master`; is a no-op on packages that
declare no extras.
- **Banned-trailer pre-merge lint**: rejects PR descriptions containing
a `Co-authored-by: ... <noreply@anthropic.com>` trailer before the PR
reaches merge, where the org ruleset would reject the squash-push
anyway. Posts a sticky comment with remediation steps; updates it to a
"resolved" state when the trailer is removed, rather than deleting
(which requires elevated token scope on fork PRs).
This commit is contained in:
Mason Daugherty
2026-05-18 15:12:21 -07:00
committed by GitHub
parent 12d5e78c3b
commit 2458a7912e
5 changed files with 717 additions and 0 deletions

View File

@@ -0,0 +1,146 @@
# Block PRs whose head ref is `main` (or `master`) from a fork. This topology
# (`<fork>:master -> langchain-ai/langchain:master`) lets contributors click
# "Update branch" on the PR, producing a `Merge branch 'master' into master`
# commit on the source side that — under admin merge override — can land
# directly on `master` as a 2-parent merge commit, bypassing the repo's
# squash-only policy and polluting the changelog.
#
# `pull_request_target` is required so the job receives a token scoped to
# write PR labels/comments on fork PRs (the standard `pull_request` token is
# read-only for forks). This also means the job MUST NOT check out PR code —
# see the inline warning in the trigger block below.
#
# Maintainer bypass: add the `bypass-fork-main-check` label to the PR.
name: Block fork main PRs
on:
pull_request_target:
# NEVER CHECK OUT UNTRUSTED CODE FROM A PR's HEAD IN A pull_request_target JOB.
# Doing so would allow attackers to execute arbitrary code in the context of your repository.
types: [opened, reopened, synchronize, labeled, unlabeled]
permissions:
contents: read
jobs:
guard:
if: >-
github.repository_owner == 'langchain-ai' &&
github.event.pull_request.head.repo.fork == true &&
(github.event.pull_request.head.ref == 'main' || github.event.pull_request.head.ref == 'master') &&
!contains(github.event.pull_request.labels.*.name, 'bypass-fork-main-check')
runs-on: ubuntu-latest
permissions:
pull-requests: write
steps:
- name: Close PR and post guidance
uses: actions/github-script@3a2844b7e9c422d3c10d287c895573f7108da1b3 # v9.0.0
with:
script: |
const { owner, repo } = context.repo;
const prNumber = context.payload.pull_request.number;
const headRef = context.payload.pull_request.head.ref;
const marker = '<!-- block-fork-main -->';
// Ensure the warning label exists and apply it
const labelName = 'fork-main-head';
try {
await github.rest.issues.getLabel({ owner, repo, name: labelName });
} catch (e) {
if (e.status !== 404) {
throw new Error(`getLabel(${labelName}) failed: ${e.message}`);
}
try {
await github.rest.issues.createLabel({
owner, repo, name: labelName, color: 'b76e79',
});
} catch (createErr) {
// A 422 with code `already_exists` means a race created the
// label between getLabel and createLabel — safe to ignore.
// Any other 422 (bad color, name too long) indicates a real
// bug introduced by editing this step, so rethrow.
const alreadyExists =
createErr.status === 422 &&
Array.isArray(createErr.errors) &&
createErr.errors.some(e => e.code === 'already_exists');
if (!alreadyExists) throw createErr;
}
}
await github.rest.issues.addLabels({
owner, repo, issue_number: prNumber, labels: [labelName],
});
const defaultBranch = context.payload.repository.default_branch;
const lines = [
marker,
`**This PR has been automatically closed** because its head branch is \`${headRef}\` on a fork.`,
'',
'PRs opened from a fork\'s `main` (or `master`) branch can produce a `Merge branch \'main\' into main` commit on the source side. Under an admin merge override that commit can land directly on this repo\'s default branch, bypassing the squash-only policy and polluting the changelog.',
'',
'To fix:',
`1. Sync your fork's \`${defaultBranch}\` first (\`git fetch upstream && git switch ${defaultBranch} && git merge --ff-only upstream/${defaultBranch}\`)`,
'2. Create a feature branch: `git switch -c feat/my-change`',
'3. Push it: `git push -u origin feat/my-change`',
`4. Open a new PR from \`feat/my-change\` → \`langchain-ai/langchain:${defaultBranch}\``,
'',
'*Maintainers: add the `bypass-fork-main-check` label to override.*',
];
const body = lines.join('\n');
// Dedup: update existing marker comment instead of stacking.
const comments = await github.paginate(
github.rest.issues.listComments,
{ owner, repo, issue_number: prNumber, per_page: 100 },
);
const existing = comments.find(c => c.body && c.body.includes(marker));
if (!existing) {
await github.rest.issues.createComment({
owner, repo, issue_number: prNumber, body,
});
} else if (existing.body !== body) {
await github.rest.issues.updateComment({
owner, repo, comment_id: existing.id, body,
});
}
if (context.payload.pull_request.state === 'open') {
await github.rest.pulls.update({
owner, repo, pull_number: prNumber, state: 'closed',
});
}
// Cancel still-queued/in-progress checks on this PR head.
// Best-effort: new runs may still queue after this loop (e.g., other
// pull_request triggers fanning out). The PR is already closed above,
// so leftover runs are wasted compute, not a correctness issue.
// We track the cancel ratio so a wholesale failure (token-scope
// regression making EVERY cancel return 403) is surfaced rather
// than silently producing N warnings + green job.
const headSha = context.payload.pull_request.head.sha;
let attempted = 0;
let cancelled = 0;
for (const status of ['in_progress', 'queued']) {
const runs = await github.paginate(
github.rest.actions.listWorkflowRunsForRepo,
{ owner, repo, head_sha: headSha, status, per_page: 100 },
);
for (const run of runs) {
if (run.id === context.runId) continue;
attempted++;
try {
await github.rest.actions.cancelWorkflowRun({
owner, repo, run_id: run.id,
});
cancelled++;
} catch (err) {
core.warning(`Could not cancel run ${run.id}: ${err.message}`);
}
}
}
if (attempted > 0 && cancelled === 0) {
core.warning(`Attempted to cancel ${attempted} run(s) on head ${headSha} but none succeeded — check token scope.`);
}
core.setFailed(`PR head ref is \`${headRef}\` on a fork — open from a feature branch instead.`);

205
.github/workflows/bump_uv_pin.yml vendored Normal file
View File

@@ -0,0 +1,205 @@
# Monthly bump of the uv pin in `.github/actions/uv_setup/action.yml`.
#
# We pin uv (rather than letting setup-uv resolve latest) because
# `releases.astral.sh` lags GitHub Releases on new uv versions, causing CI
# to flap on fresh-release days. This workflow keeps the pin fresh without
# exposing that race.
#
# Dependabot's `github-actions` ecosystem only updates `uses:` SHA pins, not
# the `UV_VERSION` env value the action passes to `astral-sh/setup-uv`, so we
# open the PR ourselves. Idempotent: if a PR for the target version already
# exists, the workflow exits without creating a duplicate.
name: "Bump uv pin"
on:
schedule:
- cron: "0 9 1 * *"
workflow_dispatch:
permissions:
contents: read
concurrency:
group: bump-uv-pin
cancel-in-progress: false
jobs:
bump:
if: github.repository_owner == 'langchain-ai'
name: "Open PR if uv has a newer release"
runs-on: ubuntu-latest
permissions:
contents: write
pull-requests: write
steps:
- uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6
- name: Resolve current and latest uv versions
id: versions
env:
GH_TOKEN: ${{ github.token }}
run: |
set -euo pipefail
action_file=".github/actions/uv_setup/action.yml"
current=$(grep -oE 'UV_VERSION: "[0-9]+\.[0-9]+\.[0-9]+"' "$action_file" \
| sed -E 's/UV_VERSION: "([^"]+)"/\1/' | head -n1)
latest=$(gh api repos/astral-sh/uv/releases/latest --jq .tag_name)
semver='^[0-9]+\.[0-9]+\.[0-9]+$'
if [[ ! "$current" =~ $semver ]]; then
echo "::error::Could not parse current uv pin from $action_file (got '$current')"
exit 1
fi
if [[ ! "$latest" =~ $semver ]]; then
echo "::error::Unexpected uv tag from GitHub API (got '$latest')"
exit 1
fi
echo "current=$current" >> "$GITHUB_OUTPUT"
echo "latest=$latest" >> "$GITHUB_OUTPUT"
echo "branch=chore/bump-uv-$latest" >> "$GITHUB_OUTPUT"
echo "Current pin: $current"
echo "Latest uv: $latest"
- name: Log if already up to date
# The actual skip is implemented by the `if:` guards on every
# subsequent step; this step only emits a log line so the run
# history shows why no PR was opened.
if: steps.versions.outputs.current == steps.versions.outputs.latest
run: echo "uv pin already at ${{ steps.versions.outputs.latest }}; nothing to do."
- name: Skip if PR already open for this version
id: existing
if: steps.versions.outputs.current != steps.versions.outputs.latest
env:
GH_TOKEN: ${{ github.token }}
BRANCH: ${{ steps.versions.outputs.branch }}
run: |
set -euo pipefail
count=$(gh pr list --head "$BRANCH" --state open --json number --jq 'length')
echo "count=$count" >> "$GITHUB_OUTPUT"
if [ "$count" -gt 0 ]; then
echo "Open PR already exists for $BRANCH; skipping."
fi
- name: Wait for astral mirror to replicate
id: mirror
if: steps.versions.outputs.current != steps.versions.outputs.latest && steps.existing.outputs.count == '0'
env:
LATEST: ${{ steps.versions.outputs.latest }}
run: |
set -euo pipefail
# The mirror can lag GitHub Releases. If it hasn't replicated yet,
# defer the bump rather than landing a pin that races the mirror
# on every CI run. We probe several arches because partial
# replication (linux ready, macOS/aarch64 not) would still race
# CI on other runners.
assets=(
"uv-x86_64-unknown-linux-gnu.tar.gz"
"uv-aarch64-unknown-linux-gnu.tar.gz"
"uv-x86_64-apple-darwin.tar.gz"
"uv-aarch64-apple-darwin.tar.gz"
)
ready=true
for asset in "${assets[@]}"; do
url="https://releases.astral.sh/github/uv/releases/download/${LATEST}/${asset}"
# `curl -sI` returns nothing on stderr at -s; capture exit code so a
# permanently broken DNS/TLS path is surfaced instead of collapsing
# to an opaque "000".
set +e
status=$(curl -sIo /dev/null -w '%{http_code}' --max-time 30 "$url" 2>/tmp/curl.err)
curl_rc=$?
set -e
echo "Mirror HEAD $url -> HTTP $status (curl exit=$curl_rc)"
if [ "$status" != "200" ]; then
ready=false
if [ "$curl_rc" -ne 0 ]; then
echo "::warning::curl failed for $asset (exit=$curl_rc): $(cat /tmp/curl.err 2>/dev/null || true)"
else
echo "::warning::astral mirror has not replicated $asset for uv $LATEST yet (HTTP $status)."
fi
fi
done
if [ "$ready" = "true" ]; then
echo "ready=true" >> "$GITHUB_OUTPUT"
else
echo "ready=false" >> "$GITHUB_OUTPUT"
echo "::warning::Deferring uv bump to $LATEST until all probed arches are mirrored."
fi
- name: Open bump PR
if: steps.versions.outputs.current != steps.versions.outputs.latest && steps.existing.outputs.count == '0' && steps.mirror.outputs.ready == 'true'
env:
GH_TOKEN: ${{ github.token }}
CURRENT: ${{ steps.versions.outputs.current }}
LATEST: ${{ steps.versions.outputs.latest }}
BRANCH: ${{ steps.versions.outputs.branch }}
DEFAULT_BRANCH: ${{ github.event.repository.default_branch }}
run: |
set -euo pipefail
action_file=".github/actions/uv_setup/action.yml"
# `grep -c` returns 1 on no-match and 2 on read errors. We want
# "no match" surfaced as the explicit count-of-zero check below;
# read errors must abort. Capture the exit code separately so
# `set -e` doesn't swallow either case.
set +e
before=$(grep -cE "UV_VERSION: \"${CURRENT}\"" "$action_file")
before_rc=$?
set -e
if [ "$before_rc" -gt 1 ]; then
echo "::error::grep read error on $action_file (exit=$before_rc)"
exit 1
fi
if [ "$before" -ne 1 ]; then
echo "::error::Expected exactly 1 'UV_VERSION: \"$CURRENT\"' in $action_file, found $before"
exit 1
fi
sed -i -E "s/UV_VERSION: \"${CURRENT}\"/UV_VERSION: \"${LATEST}\"/" "$action_file"
set +e
after=$(grep -cE "UV_VERSION: \"${LATEST}\"" "$action_file")
after_rc=$?
set -e
if [ "$after_rc" -gt 1 ]; then
echo "::error::grep read error on $action_file (exit=$after_rc)"
exit 1
fi
if [ "$after" -ne 1 ]; then
echo "::error::Expected exactly 1 'UV_VERSION: \"$LATEST\"' after sed, found $after"
exit 1
fi
if git diff --quiet "$action_file"; then
echo "No changes after sed; bailing out (current=$CURRENT, latest=$LATEST)."
exit 1
fi
# Reuse-or-recreate orphan branch from a prior run that pushed
# but failed before `gh pr create` (no open PR sits on it).
# The delete can race a concurrent run (manual workflow_dispatch
# firing while the cron is mid-flight, since concurrency group
# does not cancel-in-progress); fall through with a warning so a
# losing race does not kill an otherwise-clean job mid-state.
if git ls-remote --exit-code --heads origin "$BRANCH" >/dev/null 2>&1; then
echo "::warning::Branch $BRANCH exists on origin without an open PR; deleting before recreating."
if ! git push origin --delete "$BRANCH"; then
echo "::warning::Delete of $BRANCH failed (concurrent run, or branch already gone); the subsequent push will surface any real conflict."
fi
fi
git config --local user.name "github-actions[bot]"
git config --local user.email "41898282+github-actions[bot]@users.noreply.github.com"
git checkout -b "$BRANCH"
git add "$action_file"
git commit -m "chore(deps): bump uv to $LATEST"
git push --set-upstream origin "$BRANCH"
body_file="$(mktemp)"
{
printf 'Bumps the uv pin in `.github/actions/uv_setup/action.yml` from `%s` to [`%s`](https://github.com/astral-sh/uv/releases/tag/%s).\n\n' "$CURRENT" "$LATEST" "$LATEST"
printf 'Opened automatically by `bump_uv_pin.yml`. Mirror availability on `releases.astral.sh` was verified before this PR was created, so CI should not race the fallback.\n'
} > "$body_file"
gh pr create \
--head "$BRANCH" \
--base "$DEFAULT_BRANCH" \
--title "chore(deps): bump uv to $LATEST" \
--body-file "$body_file"

73
.github/workflows/check_extras_sync.yml vendored Normal file
View File

@@ -0,0 +1,73 @@
# See `.github/scripts/check_extras_sync.py` for the rationale.
name: "🔍 Check Extras Sync"
on:
pull_request:
paths:
- "libs/**/pyproject.toml"
- ".github/scripts/check_extras_sync.py"
- ".github/workflows/check_extras_sync.yml"
push:
branches: [master]
paths:
- "libs/**/pyproject.toml"
- ".github/scripts/check_extras_sync.py"
- ".github/workflows/check_extras_sync.yml"
concurrency:
group: ${{ github.workflow }}-${{ github.ref }}
cancel-in-progress: true
permissions:
contents: read
jobs:
check-extras-sync:
if: github.repository_owner == 'langchain-ai'
name: "Verify extras match required deps"
runs-on: ubuntu-latest
timeout-minutes: 2
steps:
- name: "📋 Checkout Code"
uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6
- name: "🐍 Set up Python and uv"
uses: "./.github/actions/uv_setup"
with:
python-version: "3.13"
enable-cache: "false"
- name: "🔍 Check extras sync"
# Iterate every package pyproject.toml under libs/. The script
# no-ops on packages without [project.optional-dependencies], so
# this is harmless on packages without extras and automatically
# picks up new partners as they're added. No `-maxdepth` cap so
# deeper future restructures (e.g. `libs/partners/<group>/<pkg>/`)
# are picked up automatically.
run: |
set -euo pipefail
mapfile -t files < <(
find libs -name pyproject.toml \
-not -path "*/.venv/*" \
-not -path "*/node_modules/*" \
-not -path "*/build/*" \
-not -path "*/dist/*" \
-not -path "*/.tox/*" \
| sort
)
if [ ${#files[@]} -eq 0 ]; then
echo "::error::No pyproject.toml files found under libs/"
exit 1
fi
failed=()
for f in "${files[@]}"; do
if ! python .github/scripts/check_extras_sync.py "$f"; then
failed+=("$f")
fi
done
if [ ${#failed[@]} -gt 0 ]; then
echo "::error::Extras-sync check failed for ${#failed[@]} package(s):"
printf '::error:: %s\n' "${failed[@]}"
exit 1
fi

175
.github/workflows/pr_lint_trailer.yml vendored Normal file
View File

@@ -0,0 +1,175 @@
# Pre-merge banned-trailer check.
name: "🏷️ PR trailer lint"
on:
pull_request:
types: [ opened, edited, synchronize, reopened ]
permissions:
pull-requests: write
jobs:
trailer-check:
if: github.repository_owner == 'langchain-ai'
name: "validate squash-merge has no banned trailers"
runs-on: ubuntu-latest
# Serialize per-PR. Rapid `edited`/`synchronize` events on a PR open can
# otherwise produce two concurrent runs that both observe "no existing
# sticky" and both call `createComment`, leaving a duplicate failure
# comment that the find-first updater will never reconcile. We queue
# (cancel-in-progress: false) rather than cancel, so the in-flight run
# finishes its sticky write before the next event evaluates.
concurrency:
group: pr-trailer-lint-${{ github.event.pull_request.number }}
cancel-in-progress: false
steps:
- name: Check PR title and body for banned trailer
uses: actions/github-script@3a2844b7e9c422d3c10d287c895573f7108da1b3 # v9.0.0
# Bound the comment-write tail so a hung GitHub API call cannot leave
# the check stuck "in progress" past the runner default. `core.setFailed`
# is invoked before the sticky write, so the failure status is already
# recorded if this timeout fires.
timeout-minutes: 5
with:
script: |
if (!context.payload.pull_request) {
core.setFailed('No pull_request payload — workflow must run on pull_request events.');
return;
}
const { title, body, number } = context.payload.pull_request;
// Normalize line endings — GitHub returns whatever the editor used,
// and CRLF leaves stray \r chars in offending-line displays.
const fullBody = (body || '').replace(/\r\n/g, '\n');
const STICKY_MARKER = '<!-- pr-trailer-lint -->';
// Mirrors the org ruleset regex on the default branch. Keep in lock-step:
// the live source of truth is the ruleset's `commit_message_pattern.pattern`
// field at GitHub org settings → Rulesets → `block-anthropic-coauthor`
// (or whichever ruleset blocks this trailer on the default branch).
// The pattern below is informational; verify against the live ruleset
// when updating either side, or this check silently passes pushes
// that the ruleset will then reject (defeating the entire purpose).
//
// Case-folding is intentionally narrow (`[Aa]`/`[Bb]`) because the
// ruleset's pattern is narrow. Do NOT add the `i` flag — that would
// catch cases the ruleset does not, surfacing false positives the
// ruleset would let through.
const BANNED_REGEX = /Co-[Aa]uthored-[Bb]y:.*<noreply@anthropic\.com>/;
const squashMessage = `${title} (#${number})\n\n${fullBody}`;
async function findStickyComment() {
const comments = await github.paginate(github.rest.issues.listComments, {
...context.repo,
issue_number: number,
per_page: 100,
});
return comments.find(c => c.body && c.body.startsWith(STICKY_MARKER));
}
// Comment write paths can fail for several reasons that should not
// turn this advisory job red on its own: fork PRs run with
// restricted tokens, secondary rate limits, transient API errors.
// Fall back to `core.summary` so a maintainer can paste the
// remediation manually. The check still fails — `setFailed` is
// invoked before this function, so the failure signal is already
// recorded by the time the comment write is attempted.
//
// The try/catch wraps ONLY the write call so that a bug in
// `findStickyComment` (e.g., pagination throwing) surfaces with
// its true cause instead of being misattributed to "fork PR token".
async function postStickyOrSummary(commentBody, summaryHeading) {
const existing = await findStickyComment();
try {
if (existing) {
if (existing.body !== commentBody) {
await github.rest.issues.updateComment({
...context.repo,
comment_id: existing.id,
body: commentBody,
});
}
} else {
await github.rest.issues.createComment({
...context.repo,
issue_number: number,
body: commentBody,
});
}
} catch (commentErr) {
core.warning(`Could not post sticky comment (fork PR token, rate limit, or transient API error): ${commentErr.message}`);
await core.summary
.addHeading(summaryHeading)
.addRaw('Paste the following into the PR as a comment:')
.addCodeBlock(commentBody, 'markdown')
.write();
}
}
const lines = squashMessage.split('\n');
const offendingIndices = [];
for (let i = 0; i < lines.length; i++) {
if (BANNED_REGEX.test(lines[i])) {
offendingIndices.push(i);
}
}
if (offendingIndices.length === 0) {
core.info('No banned trailer in squash-merge message.');
// Mark any prior failure comment as resolved. We update rather
// than delete because `deleteComment` 403s under restricted
// fork-PR tokens, whereas `updateComment` on a bot-authored
// comment works in both modes. Wrapped in try/catch because a
// transient API failure during cleanup must NOT turn a green
// check into red.
try {
const existing = await findStickyComment();
if (existing) {
const resolvedBody = [
STICKY_MARKER,
'✅ **Trailer fixed.** The previous warning is resolved.',
].join('\n');
if (existing.body !== resolvedBody) {
await github.rest.issues.updateComment({
...context.repo,
comment_id: existing.id,
body: resolvedBody,
});
}
}
} catch (cleanupErr) {
core.warning(`Check passed but could not update prior failure comment to resolved: ${cleanupErr.message}`);
}
return;
}
const offendingExcerpt = offendingIndices
.map(i => `Line ${i + 1}: ${lines[i]}`)
.join('\n');
const commentBody = [
STICKY_MARKER,
'⚠️ **Banned trailer in PR — would block the squash-merge push to the default branch.**',
'',
'The would-be squash-merge commit message contains a `Co-authored-by: ... <noreply@anthropic.com>` line. An organization ruleset on the default branch rejects any push whose commit message matches that pattern, so this PR cannot be merged until the trailer is removed.',
'',
'**Found:**',
'```',
offendingExcerpt,
'```',
'',
'### Fix',
'',
'Edit the PR description and remove the offending line(s). The trailer is auto-inserted by some Claude-based authoring tools — strip it before opening or merging the PR. Save the description; this check will re-run automatically.',
].join('\n');
// Set the failure signal BEFORE the sticky write — if the comment
// API hangs, the runner-level timeout fires with the failure
// status already recorded. Reversing the order leaves the check
// stuck "in progress" instead of red.
core.setFailed(`PR contains banned trailer matching ${BANNED_REGEX}`);
await postStickyOrSummary(
commentBody,
'Banned trailer in PR; comment could not be posted',
);