Skip to content

fix(apps): pin md-notebook at app sdk level 10 - #38669

Merged
bolichen97 merged 1 commit into
mainfrom
fix/repin-md-notebook-level-10
Nov 17, 2026
Merged

bolichen97 merged 1 commit into
mainfrom
fix/repin-md-notebook-level-10

Conversation

@bolichen97

Copy link
Copy Markdown
Contributor

Problem / Motivation

Goal: The md-notebook adoption pin verifies against the level its credentialMasks needs.

python3 scripts/adoption_pin.py verify refuses the md-notebook row on main while every other row passes:

md-notebook: minAppSdkLevel 8 is below App Kit seam level 10, which the pinned tree uses (app.json credentialMasks tool gate); compute it with --min-sdk-level 10 or higher

The row was computed at level 8 shortly before #37459 merged. #37459 made the agent's file tools fence an installed app's credentialMasks, so adoption_pin.py counts any credentialMasks declaration as a level-10 use. md-notebook's pinned app.json declares "credentialMasks": ["state/"]. Reported in #38443 and #38478.

Why it matters

The pin table is the release captain's source of truth for adoption, and verify stops at its first refusal, so every later pin PR has to measure the remaining rows one at a time around this one. The row also states a level below what the pinned tree relies on: on a build at level 8 or 9 the adopted tree's state/ mask would be held by the OS sandbox only, not by the agent's file tools.

Not a goal

  • Changing any other pin row, the md-notebook repository, its pinned commit or the wheel copy.
  • Running adoption_pin.py verify in CI, or any change to how seam levels are computed.

What changed (motivation → approach → change)

Symptom: verify refuses one row. Cause: the row's minAppSdkLevel predates the level #37459 assigned to credentialMasks. Change: scripts/adoption_pin.py compute https://github.com/KiroCrewCaptain/kirocrew-app-md-notebook ad86662fc977fa3c459a4477021d9076d000d650 --min-sdk-level 10 prints a row that differs from the committed one only in minAppSdkLevel (8 to 10); the commit, version, treeSha256 and archiveSha256 are identical. That one field changes in:

  • src/kiro_crew/apps/adoption-pins.json
  • _SHIPPED_PINS in test/test_builtin_adoption_pins.py
  • docs/system-specs/modules/md-notebook.md, whose adoption paragraph stated the pin's level as 8. The State Layout line calling the state/ directory entry a level-8 seam stays, since that is the level of the directory-entry form itself.

App repo: KiroCrewCaptain/kirocrew-app-md-notebook@ad86662

Wheel-copy base: md-notebook@d3b30f159e924f1d9f35b964881222ee205babe3

That is #37572's merge commit. #37572 named f88806d as the port base; the only wheel-copy change between the two is #37572's own app.json version move to the pinned 1.0.2 (D12), which the pinned tree already carries, and the wheel copy has not changed since.

Backwards compatibility

Compatible: the pin's minimum level is compared with APP_SDK_LEVEL, a constant of the same wheel that ships the pin table, and this wheel's level is 10, so a host running it adopts md-notebook exactly as before.

Who reads minAppSdkLevel:

  • builtin_adoption_pins.parse_adoption_pins accepts any integer of at least 1, so 10 parses (test_the_shipped_pin_file_validates_and_pins_exactly_the_shipped_table).
  • builtin_adoption_checks.needs_newer_sdk (E5, called from builtin_adoption.py and builtin_adoption_apply.py) compares 10 with APP_SDK_LEVEL 10, so it stays False and STAGE and APPLY run as before.
  • scripts/adoption_pin.py verify (the release captain) now reports md-notebook ok, and every row ok.

Removes nothing: no row, key or file is removed; one row's minimum level moves up to the level the shipping build already has.

Tests

_SHIPPED_PINS in test/test_builtin_adoption_pins.py moves to 10 with the table, so test_the_shipped_pin_file_validates_and_pins_exactly_the_shipped_table keeps pinning the shipped file byte for byte.

Proof: PROVEN test/test_builtin_adoption_pins.py::test_the_shipped_pin_file_validates_and_pins_exactly_the_shipped_table (head 4dc1d84)

Run serially (-n0, isolated --basetemp): test_builtin_adoption_pins.py, test_legacy_state_moves.py, test_builtin_adoption_gate.py, test_builtin_adoption_release.py, test_adopted_wheel_copy_check.py, test_app_agent_tools_adopted.py, test_md_notebook.py (364 passed); integration/test_builtin_adoption_boot.py, test_md_notebook_state.py, test_core_md_notebook_decoupling.py (67 passed, 2 skipped).

Manual verification

python3 scripts/adoption_pin.py verify over the network: refuses md-notebook on main; on this head every row reports ok (agent-worlds, auto-improvement, auto-research, aws-control, code-review-sage, design-critique, design-tweak, dev-fleet, file-explorer, issue-radar, md-notebook, meetings, ops-mission-control, papyrus, personal-shopper, pptx-maker, project-scaffolder, spec-builder).

Related Issues

Part of #26967

Sibling sweep

Swept: python3 scripts/adoption_pin.py verify, which measures every row of adoption-pins.json against the seam levels its pinned tree uses
Siblings: none -- verify reports every other row ok at its pinned level

Pattern harvest

Rule candidate: review-prompt
Pattern: a change that raises the App Kit seam level of an existing manifest key or export leaves already pinned rows below it; such a change re-runs adoption_pin.py verify over every row.

Checklist

  • At most two commits (one is the norm), with a Conventional Commits title (feat|fix|docs|style|refactor|perf|test|chore|ci|build|revert: ...)
  • Existing tests pass and new tests added for new functionality
  • Self-review completed; code follows project style guidelines
  • Documentation updated (if applicable)
  • No secrets, credentials, or internal references in the diff

Contribution License Agreement

Contributed under the project's Apache-2.0 license.

The md-notebook adoption row was computed at minAppSdkLevel 8 just
before #37459 made any credentialMasks declaration a level-10 use, so
scripts/adoption_pin.py verify now refuses that row. Recomputed at
the same commit with --min-sdk-level 10: only the level changes; the
commit, version and both digests are identical. The _SHIPPED_PINS
table and the md-notebook spec carry the same level.

Part of #26967
@bolichen97
bolichen97 requested review from a team and KiroCrewCaptain as code owners November 17, 2026 17:29
@bolichen97
bolichen97 enabled auto-merge (squash) November 17, 2026 17:29
@github-actions github-actions Bot added the readiness: checking Automated validation is still running label Nov 17, 2026
@github-actions

Copy link
Copy Markdown
Contributor

UX Review (Opus 5.5) — ⏭️ skipped

Revision 4dc1d8452a22e297a02343eef3b35ad3d5b01377 touches no user-facing surface (no changes under website/ or committed screenshots), so the UX review was skipped. Does not block merge.

@github-actions

Copy link
Copy Markdown
Contributor

Design Review (Opus 5.5) — ✅ PASS

Design-level review of 4dc1d8452a22e297a02343eef3b35ad3d5b01377 — updated in place on each push. A BLOCK verdict blocks PR readiness; PASS does not.

Design-Verdict: PASS

Nothing blocks: this is a data correction that brings one pin row up to the seam level its pinned tree uses. The shipping wheel's APP_SDK_LEVEL is already 10, so adoption behaves the same.

[DESIGN-REVIEWED] 4dc1d84

Description read

[DESCRIPTION-READ] 8b850ee781f5127d2d0d2f34961de9e9fef12c3d811f606797d619b063f0b8f5

That sha256 names everything this verdict read: the PR description, plus the 1 evidence file(s) the workflow collected from it and handed this review. It covers that evidence because the media strip replaces every attachment URL with the same placeholder, so an edit swapping one attachment for another would leave the description's captured text identical while what this review judged changed. Recompute it with .github/scripts/pr-description-capture.sh over the same inputs; if it differs, the description or that evidence changed after this verdict was formed and any finding drawn from either is unproven.

@github-actions

Copy link
Copy Markdown
Contributor

First Principles Review (Opus 5.5) — ✅ PASS

Premise-level review of 4dc1d8452a22e297a02343eef3b35ad3d5b01377 — why this exists and whether the shipped surface is the smallest honest version. Updated in place on each push. A BLOCK verdict blocks PR readiness; PASS does not.

First-Principles-Verdict: PASS

Nothing to check.

What this change ships

Inventory (3 items) — 3 justified

Intent: let the release captain's adoption_pin.py verify accept the md-notebook pin by stating the seam level its pinned tree actually uses. This is a FIX.

  1. The md-notebook adoption pin now declares a minimum App SDK level of 10 instead of 8 — justified
  2. The shipped-pin test table is updated to match, so the test keeps pinning the file byte for byte — justified
  3. The md-notebook spec's adoption paragraph now gives 10 as the pin level, and says why — justified

[FIRST-PRINCIPLES-REVIEWED] 4dc1d84

Description read

[DESCRIPTION-READ] 6fa65e5b53dc99dbd9c090b0bd0a070eb2cd8fa92c89cca93eb767306482f089

That sha256 names the PR description this verdict read. Recompute it with .github/scripts/pr-description-capture.sh; if it differs, the description changed after this verdict was formed and any finding drawn from the description is unproven.

@github-actions

Copy link
Copy Markdown
Contributor

Opus 5.5 Review — ✅ no blocking findings

Reviewed 4dc1d8452a22e297a02343eef3b35ad3d5b01377 — this comment is updated in place on each push.

Review details

No findings.

[OPUS-REVIEWED] 4dc1d84

Verdict parsed from the review's SHA-scoped output markers for commit 4dc1d8452a22e297a02343eef3b35ad3d5b01377.

False positive or not applicable? A repository writer can comment:
/ai-review override fable 4dc1d8452a22e297a02343eef3b35ad3d5b01377 false-positive: <reason, 20 to 500 characters>

@github-actions

Copy link
Copy Markdown
Contributor

GPT 6.1 Review — ✅ no blocking findings

GPT 6.1 completed its review of 4dc1d8452a22e297a02343eef3b35ad3d5b01377 and found no blocking issues.

This comment is updated in place on each push.

Review details

No findings.
[GPT-REVIEWED] 4dc1d84

False positive or not applicable? A repository writer can comment:
/ai-review override gpt 4dc1d8452a22e297a02343eef3b35ad3d5b01377 false-positive: <reason, 20 to 500 characters>

@KiroCrewCaptain KiroCrewCaptain left a comment

Copy link
Copy Markdown
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Checked head 4dc1d84. The only data change is md-notebook's minAppSdkLevel moving from 8 to 10, in both the shipped adoption pin table and the test's pinned copy of it; commit, version, tree digest and archive digest are unchanged. Running the pin compute tool against the pinned app commit at level 10 reproduces the committed row exactly, and the same run at level 9 is refused because the app declares credentialMasks, which is a level 10 seam. This wheel's App SDK level is already 10, so the needs-newer-SDK check stays false and adoption on a running host behaves as before. The adoption pin, adoption gate and md-notebook test files pass on this head, and nothing blocks.

@github-actions github-actions Bot added readiness: passed Eligible automated validation passed for the current revision and removed readiness: checking Automated validation is still running labels Nov 17, 2026
@bolichen97
bolichen97 merged commit ea5fa49 into main Nov 17, 2026
99 checks passed
@bolichen97
bolichen97 deleted the fix/repin-md-notebook-level-10 branch November 17, 2026 18:29
@github-actions github-actions Bot removed the readiness: passed Eligible automated validation passed for the current revision label Nov 17, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants