fix: #420 fix(deps): bump the babel group across 1 directory with 5 updates

[ianwremmel-ai-agent]

Automated fix branch for #420 (fix(deps): bump the babel group across 1 directory with 5 updates).

Why #420 is red

Head efe16a55d102e4d8a05b80edf07b1f555f3cfc88 on dependabot/npm_and_yarn/babel-ae2feaba14. Failing:

What this PR is

This PR targets the Dependabot branch, not the default branch: merging it re-runs CI on #420, which the automation then merges once green. Opened by the coven Dependabot automation (coven/infrastructure/services/dependabot).

[ianwremmel-ai-agent]

Plan

State as of c381f41 (Dependabot force-pushed #420's branch at 2026-09-08 12:54
UTC). Everything this plan said before that rebase — the check-run-reporter
test-splitting blocker — was read off the old head and has been dropped.

• Read fix(deps): bump the babel group across 1 directory with 5 updates #420's failure on run
  34228744849.
  build dies in npm ci with ERESOLVE: the group took @babel/core to
  ^8.0.1 and @babel/preset-env to ^8.0.2 but left
  @babel/preset-typescript at ^7.24.6 (peer @babel/core@^7.0.0-0).
  Every other check is skipped.
• Built the obvious completion (@babel/preset-typescript@^8) and confirmed
  it makes npm ci pass.
• Push that completion. Dropped. babel.config.js exists only for jest,
  jest.config.ts sets no transform, and babel-jest@29.7.0 peers
  @babel/core@^7.8.0, so jest transforms with a nested Babel 7 while the
  root presets are Babel 8. 22 of 202 source files then compile to invalid
  JavaScript (0 of 202 with preset-typescript 7). It trades a loud failure
  for a silent one, so this branch stays one empty commit.
• Ruled out an npm overrides workaround: @jest/transform@29 calls
  loadPartialConfig synchronously (callback-only in Babel 8) and
  babel-preset-jest@29 passes the removed blacklist option.
• Identified the prerequisite: babel-jest@30.5.1 peers
  @babel/core: ^7.11.0 || ^8.0.0-0, so jest 30 must land first. fix(deps): bump the jest group across 1 directory with 6 updates #442
  (fix(deps): bump the jest group with 6 updates) is that upgrade, opened
  in the same 12:54 batch. It is red on test-unit for an unrelated reason
  — jest 30's TS config loader rejects import pkg from './package.json'
  without an import attribute.
• Checked main: run
  34182043796
  fails diff, test-unit and all five test-integration shards, so a
  fully green fix(deps): bump the babel group across 1 directory with 5 updates #420 is not reachable from this branch either way.
• Consolidated every earlier comment into the single diagnosis comment and
  marked the superseded ones.
• Blocked on an operator scope decision (section 5 of the diagnosis): close
  fix(deps): bump the babel group across 1 directory with 5 updates #420 and ignore the babel majors, fix fix(deps): bump the jest group across 1 directory with 6 updates #442 first, or extend this branch
  with the jest 29 -> 30 upgrade.

No dependency version has been changed on this branch.

[ianwremmel-ai-agent]

Diagnosis: #420 cannot install, and cannot work until jest 30 lands

Current as of c381f41, the head Dependabot force-pushed at 12:54 UTC on
2026-09-08. This replaces every earlier analysis on this PR — the
check-run-reporter story in the older comments was read off the pre-rebase head
and no longer applies.

1. npm ci fails, so every other check is skipped

Run 34228744849
fails in build; diff, lint, test-unit, test-integration and release
never start:

npm error code ERESOLVE
npm error While resolving: @babel/preset-typescript@7.24.6
npm error Found: @babel/core@8.0.1
npm error Could not resolve dependency:
npm error peer @babel/core@"^7.0.0-0" from @babel/preset-typescript@7.24.6

The group bump took @babel/core to ^8.0.1 and @babel/preset-env to
^8.0.2 but left @babel/preset-typescript at ^7.24.6. As composed, the
bump does not install.

2. Completing the bump does not give a working test run on jest 29

@babel/preset-typescript@8.0.1 peers @babel/core@^8.0.0, so raising it
clears the ERESOLVE. It does not get the tests running. babel.config.js
exists only for jest — it asserts NODE_ENV === 'test' with "Babel should
only be used for jest" — jest.config.ts sets no transform, so babel-jest
is the transform, and babel-jest@29.7.0 peers @babel/core@^7.8.0. npm
therefore nests a Babel 7 core under jest while the root presets are Babel 8.
Driving that configuration through the real jest pipeline in an earlier local
run compiled 22 of 202 source files to invalid JavaScript (0 of 202 with
preset-typescript 7), because Babel 8 renamed AST fields that core 7 does not
produce:

FAIL packages/@code-like-a-carpenter/exception/src/exception.test.ts
  class Subclass extends _exception.Exception<object> {}
                                             ^
  SyntaxError: Unexpected token '<'

An npm overrides entry forcing Babel 8 on jest fails differently:
@jest/transform@29 calls loadPartialConfig synchronously (Babel 8 made it
callback-only), and babel-preset-jest@29 passes the removed blacklist
visitor option.

So there is no version-preserving change on this branch that turns #420 green.

3. The prerequisite is #442, which is red for an unrelated reason

babel-jest@30.5.1 peers @babel/core: ^7.11.0 || ^8.0.0-0, which lets npm
hoist a single Babel 8. #442 (fix(deps): bump the jest group with 6 updates,
jest 29.7.0 -> 30.5.1) is that upgrade, opened in the same 12:54 batch as this
rebase. Its build, lint, TSC and diff pass; test-unit fails on
something that has nothing to do with babel:

Error: Jest: Failed to parse the TypeScript config file .../jest.config.ts
  TypeError [ERR_IMPORT_ATTRIBUTE_MISSING]: Module ".../package.json" needs an
  import attribute of "type: json"

That is jest 30's native TS config loader rejecting
import pkg from './package.json' in jest.config.ts. It needs a fix on #442's
branch.

4. Main is red on its own

Even a correct babel fix cannot produce a fully green #420 right now. Main's
latest run
(34182043796,
3bff384) fails diff, test-unit and all five test-integration shards.
test-unit there is ts-node compiling jest.config.ts without
esModuleInterop or resolveJsonModule (TS1259, TS2732); a fix is in
flight on clc-15-update-workbench-nx (5ab96c4).

5. What I need from you

This branch is still one empty commit and no dependency version has been
touched. Pick one:

1. Close fix(deps): bump the babel group across 1 directory with 5 updates #420 and add a dependabot ignore for the babel group's majors
   until jest 30 has landed. Dependabot re-raises the bump afterwards. Smallest
   and safest; I would close this PR with it.
2. Fix fix(deps): bump the jest group across 1 directory with 6 updates #442 first — the jest.config.ts import attribute — let jest 30
   merge, then come back here. fix(deps): bump the babel group across 1 directory with 5 updates #420 would still need
   @babel/preset-typescript raised to ^8 on top, which I would push on this
   branch at that point.
3. Let me do all of it on this branch: jest 29 -> 30 plus the
   preset-typescript completion. That also drags in
   packages/@code-like-a-carpenter/tool-inliner, which pins
   @babel/generator, @babel/template and @babel/types at ^7.24.6, and
   turns a dependency bump into a test-framework major.

I am not picking between these unilaterally. Parked under
dependabot-fix-parked until you reply; a comment, review or push here
requeues me.

Caveat on the review gate

Codex is unauthenticated in this pod (401 from api.openai.com), so the
cross-family adversarial review that would precede a push ran degraded on the
earlier attempt — same-family reviewers instead. Nothing was pushed, so nothing
reached CI unreviewed.

[ianwremmel-ai-agent]

Superseded. Written before Dependabot's 2026-09-08 12:54 UTC rebase; the conclusion here is wrong. See the diagnosis comment.

Resume check 2026-09-08: still blocked, and main has picked up two more failures

Re-checked the blocker from the diagnosis above. It is unchanged.

test-integration still cannot start on any branch. On main's latest run
34182043796
(3bff384) all five shards die the same way, before any test runs:

./scripts/crr-sam: line 35: testfile: unbound variable

Still downstream of check-run-reporter's test-splitting request returning 404,
which leaves steps.split.outputs.tests empty.

New since the diagnosis, both on main and both unrelated to this branch:

• test-unit — jest cannot parse jest.config.ts: TS1259: Module '"node:assert"' can only be default-imported using the 'esModuleInterop' flag, the same for node:path, plus TS2732: Cannot find module './package.json'.
• diff — The CI build process altered 1 tracked files. Please run build and commit the changes.

One correction to the record above. The checks showing green on this PR are only
the two Socket Security ones. The Push workflow has never run on this branch at
all — gh run list --branch dependabot-fix/pr-420 returns nothing, because the
head commit is chore: open PR [skip ci]. So there is no positive evidence that
the babel bump passes; there is only the evidence in the diagnosis that the bump
cannot be what breaks #420.

What I need from you

Nothing on this branch moves until check-run-reporter's test-splitting API is
restored for this repo. Once it is, say so here and I will push to this branch to
get a real run.

The two new main failures are separate from this PR. If you want them fixed, say
so and I will file them as their own work.

Still parked under dependabot-fix-parked. No dependency version was touched.

[ianwremmel-ai-agent]

Superseded. Written before Dependabot's 2026-09-08 12:54 UTC rebase; the conclusion here is wrong. See the diagnosis comment.

Resume check 2026-09-08 (third pass): blocker unchanged, but it now has an owner

Re-verified against the freshest run in the repo — 34194980420 on clc-1137-…, 06:30 UTC. All five test-integration shards still die the same way, before any test runs:

##[group]Sending 5 test names to Check Run Reporter
Check Run Reporter return a 404
./scripts/crr-sam: line 35: testfile: unbound variable

New since the last check: #437 (fix(ci): shard example tests in-repo instead of via check-run-reporter) replaces that split step. The blocker is therefore no longer an external API someone has to restore; it is #437, which is still draft and whose own integration shards are still red.

Where that leaves this PR

This branch is one empty commit. git diff against dependabot/npm_and_yarn/babel-ae2feaba14 is empty, and per the diagnosis above it should stay empty — the babel bump does not cause #420's failures.

Merging this PR is still the lever that re-runs CI on #420. The order that works:

1. fix(ci): shard example tests in-repo instead of via check-run-reporter #437 lands, so test-integration can start again.
2. Merge this PR into fix(deps): bump the babel group across 1 directory with 5 updates #420's branch to trigger a real run.
3. If fix(deps): bump the babel group across 1 directory with 5 updates #420 goes green, the automation merges it. If the aws-otel 403s come back, that is the separate examples-pipeline bug, not babel.

What I need from you

Approve this PR, or reply "go" here, and I will merge it once #437 has landed. Step 2 needs your approval and I cannot take it without one. This is the third check with no reply on the previous two, so I am recording the item as blocked on you rather than polling again.

Still parked under dependabot-fix-parked. No dependency version was touched.

[ianwremmel-ai-agent]

@ianwremmel this PR needs a scope decision from you.

Dependabot's babel group bump (#420) cannot install: it took @babel/core to 8
but left @babel/preset-typescript at 7. Completing it needs jest 30 (#442)
first, because babel-jest@29 pins jest to Babel 7. Options are in section 5 of
the diagnosis comment:
close #420 and ignore the babel majors for now, fix #442 first, or let me do
the jest 30 upgrade on this branch.

Parked until you reply. A comment, review or push here requeues this PR.

[ianwremmel-ai-agent]

Folded into the diagnosis comment,
which is the one place this PR's conclusion lives. Short version: Dependabot's
12:54 UTC rebase changed the picture — the babel bump does break the build, my
first four comments here were read off a stale head, and the decision I need is
in section 5 of that comment.

Done.