Skip to content

chore(main): release 7.10.0 - #964

Merged
devseunggwan merged 1 commit into
mainfrom
release-please--branches--main
Aug 16, 2026
Merged

devseunggwan merged 1 commit into
mainfrom
release-please--branches--main

Conversation

@devseunggwan

@devseunggwan devseunggwan commented Aug 12, 2026 •

Copy link
Copy Markdown
Owner

🤖 I have created a release beep boop

7.10.0 (2026-08-16)

Added

  • cmux-delegate: add a worker decision gate (0087b84)
  • cmux-delegate: classify worker liveness (3614663)
  • hook: demote git-commit to advise, add ADVISE-channel arm (#1030) (df6cc4e)
  • hook: re-ask before a mutation the user already refused (#1028) (f4182c1)
  • hooks: add Bash blast-radius ask to output-block-falsify-advisory (#1025) (6aa725b)
  • hooks: advise commit decomposition from the message itself (#977) (d65c1b2)
  • hooks: advise on foreground Bash calls declaring long timeouts (#1026) (83a95be)
  • hooks: advisory for n=1 quantitative claims (#969) (09181e8)
  • hooks: enumerate required tokens on block (512fd69), closes #941
  • hooks: gate poll-loop retries on reading the guard's own spec (#1027) (c4bd6dc)
  • hooks: require a non-mutating tier in approval menus (#966) (3efa1ed)
  • retrospect: add denied-actions lane + Gate-12 (#1029) (80110bd)
  • skill: add merge-briefing procedure (#980) (635f561)
  • skills: spec-drift and a spec store outside the repo (#1006) (ea24d15)

Fixed

  • cmux-delegate: key the report on the workspace (6d8d653)
  • cmux-delegate: supply the stdin column's non-TTY exemption (#1022) (615d27e), closes #981
  • codex-review-wrap: always background the review call (#983) (e18de75)
  • hook: make the anchor gate's PostToolUse findings reach the model (#1000) (7262740)
  • hooks: bind each body flag to its own gh invocation in perf advisory (#1015) (2eefdfc), closes #973 #973
  • hooks: fold lines across an open quote in safe_tokenize (#1014) (c79d872)
  • hooks: guard session state read-modify-write (#965) (5efc14c)
  • hooks: merge gates fire on non-merge commands (#986) (8dfafdf), closes #985
  • hooks: serialize jq-config dedup state and stage per process (#1017) (0e755d1), closes #970 #951
  • hooks: stop rule 1b firing on refusal-led turns (#976) (f3979aa)
  • menu-tier: gate create/update on shared surface, cancel on mutation (#1016) (8a62c97)
  • retrospect: escalate own-org public repo writes at Gate-4 (#1024) (256b428)
  • spec-drift: run Verify commands with stdin at /dev/null (#1018) (487aea8), closes #1008

Changed

  • architecture: state stdin column precondition (1a5ea64)
  • bump codeql-action/analyze to 4.37.6 (ddea2e3)
  • bump codeql-action/init to 4.37.6 (e1acc9a)
  • bump reviewdog/action-actionlint from 1.73.0 to 1.73.1 (#1032) (075ac78)
  • codex-review-wrap: pin Step 4b and Liveness to companion 1.0.6 (#1020) (cb0b375)
  • ethos: forbid delegating a hook workaround to the user (#1023) (ad1ecb5)
  • retrospect: declare memory-lint CI split intended, not a gap (#1021) (0a096ae), closes #975
  • spec: a Verify oracle must fail when its inputs are gone (#1019) (f66836e), closes #1011 #1008
  • spec: tracked feature-spec convention under .praxis/specs (#1002) (7821fa8)

This PR was generated with Release Please. See documentation.

@coderabbitai

coderabbitai Bot commented Aug 12, 2026 •

Copy link
Copy Markdown

Review Change Stack

Note

Reviews paused

It looks like this branch is under active development. To avoid overwhelming you with review comments due to an influx of new commits, CodeRabbit has automatically paused this review. You can configure this behavior by changing the reviews.auto_review.auto_pause_after_reviewed_commits setting.

Use the following commands to manage reviews:

  • @coderabbitai resume to resume automatic reviews.
  • @coderabbitai review to trigger a single review.

Use the checkboxes below for quick actions:

  • ▶️ Resume reviews
  • 🔍 Trigger review

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info
⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: fc8fbe74-efe7-4f1e-8dbe-ef0cd2926a17

📥 Commits

Reviewing files that changed from the base of the PR and between 4302f4f and ae25574.

📒 Files selected for processing (1)
  • CHANGELOG.md
🚧 Files skipped from review as they are similar to previous changes (1)
  • CHANGELOG.md

📝 Walkthrough

Walkthrough

This release updates project, marketplace, plugin, extension, and release manifest versions from 7.9.1 to 7.10.0. It adds the 7.10.0 changelog entry dated 2026-08-15 with feature additions, fixes, and process changes.

Changes

Release metadata

Layer / File(s) Summary
Synchronized version metadata
.agents/plugins/marketplace.json, .claude-plugin/..., .cursor-plugin/plugin.json, .opencode/plugin.json, .release-please-manifest.json, VERSION, gemini-extension.json, plugins/praxis/.codex-plugin/plugin.json
Updates project, marketplace, extension, and plugin versions from 7.9.1 to 7.10.0.
Release changelog entry
CHANGELOG.md
Adds the 7.10.0 entry dated 2026-08-15. It documents new hooks, skills, merge briefing, and spec-drift support; fixes for workspace reporting, review execution, findings delivery, session-state updates, merge-gate behavior, and Verify commands; and updated release conventions.

Estimated code review effort: 1 (Trivial) | ~3 minutes

Merge Risk: ⚪ Minimal · up to ae255

This release updates version metadata and changelog information, with no actionable merge-blocking risk remaining beyond normal checks and review.

Possibly related PRs

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 inconclusive)

Check name Status Explanation Resolution
Linked Issues check ❓ Inconclusive The summaries show release metadata and changelog updates, but provide no evidence that the linked issue implementations are included. Provide implementation or test evidence for the requirements in issues #941, #973, #985, #1008, and #1011.
✅ Passed checks (4 passed)
Check name Status Explanation
Out of Scope Changes check ✅ Passed The changes are limited to consistent version metadata and documentation for the stated 7.10.0 release.
Docstring Coverage ✅ Passed No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Description Check ✅ Passed Check skipped - CodeRabbit’s high-level summary is enabled.
Title check ✅ Passed The title clearly identifies this pull request as the 7.10.0 release, which matches the version updates and changelog changes.
✨ Finishing Touches
🧪 Generate unit tests (beta)
  • Create PR with unit tests
  • Commit unit tests in branch release-please--branches--main

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

Comment @coderabbitai help to get the list of available commands.

@devseunggwan
devseunggwan force-pushed the release-please--branches--main branch from 5a7c3cc to 806a48b Compare August 12, 2026 06:50
@devseunggwan devseunggwan changed the title chore(main): release 7.9.2 chore(main): release 7.10.0 Aug 12, 2026
@devseunggwan
devseunggwan force-pushed the release-please--branches--main branch 7 times, most recently from f47577b to 99b9b2f Compare August 13, 2026 09:28
@devseunggwan
devseunggwan force-pushed the release-please--branches--main branch 17 times, most recently from 245d4b7 to 1172825 Compare August 15, 2026 13:30
@devseunggwan
devseunggwan force-pushed the release-please--branches--main branch 10 times, most recently from 5814504 to 5060e45 Compare August 16, 2026 11:56
@devseunggwan
devseunggwan force-pushed the release-please--branches--main branch from 5060e45 to 01832ec Compare August 16, 2026 12:48
@devseunggwan

devseunggwan commented Aug 16, 2026 •

Copy link
Copy Markdown
Owner Author

Verification — 01832ec (rev 2)

# Claim Result
1 버전이 9개 매니페스트 전부에서 7.10.0 으로 일치한다 PASS(live)
2 changelog 가 v7.9.1..main 을 설정된 정책대로 빠짐없이 담고, 실재하지 않는 커밋은 0건이다 PASS(live)
3 이 PR 은 런타임 경로를 전혀 건드리지 않는다 — 버전·changelog 파일뿐 PASS(live)
4 릴리스될 내용이 CI 진입점 전량을 통과한다 PASS(live)
5 머지 후 태그·릴리스가 만들어지고, 배포 표면이 7.10.0 을 내놓는다 PASS(live)
Unverified — (a) 설치된 플러그인이 실제로 기동하는지는 클라이언트 쪽 관측이라 여기서 측정 불가, (b) memory frontmatter lint 단계는 로컬 사유로 실패

(a) 행 5 는 배포 표면이 무엇을 내놓는가까지 확인합니다 — 각 클라이언트가 그것을 받아 훅을 실제로 기동하는지는 클라이언트 쪽에서만 관측되고, 이 저장소에서 낼 수 있는 관측이 아닙니다. 남은 구멍은 "제공된다"와 "동작한다" 사이입니다.

(b) CI 진입점 실행에서 memory frontmatter lint 단계만 실패합니다. 대상은 저장소 밖 로컬 개인 memory 디렉터리이고, CI 에는 그 디렉터리가 없어 스킵됩니다. 이 브랜치와 무관하며 gh pr checks 964 의 test 잡이 pass 인 것이 그 대조입니다(Evidence 4).

Evidence 1 — 9개 매니페스트를 같은 프로브로 재고, `main` 을 양성 대조로 둠

버전 불일치는 플랫폼 하나만 옛 버전에 남아도 그 플랫폼 사용자에게만 조용히 드러나므로, 파일 하나씩이 아니라 릴리스가 건드리는 전량을 같은 프로브로 훑습니다. 패턴을 7\. 로 좁히면 다른 계열 버전을 놓치므로 semver 전체를 잡습니다.

$ for f in VERSION .release-please-manifest.json .claude-plugin/plugin.json .cursor-plugin/plugin.json .opencode/plugin.json gemini-extension.json plugins/praxis/.codex-plugin/plugin.json .claude-plugin/marketplace.json .agents/plugins/marketplace.json; do printf '%-45s ' "$f"; git show FETCH_HEAD:"$f" | grep -oE '[0-9]+\.[0-9]+\.[0-9]+' | sort -u | tr '\n' ' '; echo; done
VERSION                                       7.10.0
.release-please-manifest.json                 7.10.0
.claude-plugin/plugin.json                    7.10.0
.cursor-plugin/plugin.json                    7.10.0
.opencode/plugin.json                         7.10.0
gemini-extension.json                         7.10.0
plugins/praxis/.codex-plugin/plugin.json      7.10.0
.claude-plugin/marketplace.json               7.10.0
.agents/plugins/marketplace.json              7.10.0

sort -u 를 거치므로 한 파일이 두 버전을 담고 있으면 그 줄에 둘 다 찍힙니다. 아홉 줄 모두 값이 하나입니다.

Positive control: 같은 프로브를 릴리스 전 main 에 걸면 다른 값이 나옵니다 — 이것이 없으면 위 출력이 "일치" 인지 "프로브가 항상 같은 값을 찍는" 것인지 구분되지 않습니다.

$ for f in VERSION .claude-plugin/plugin.json; do printf '%-45s ' "$f"; git show origin/main:"$f" | grep -oE '[0-9]+\.[0-9]+\.[0-9]+' | sort -u | tr '\n' ' '; echo; done
VERSION                                       7.9.1
.claude-plugin/plugin.json                    7.9.1

FETCH_HEAD 는 git fetch origin release-please--branches--main 로 가져온 이 PR 의 head 입니다.

Evidence 2 — changelog 가 나열한 커밋을 실제 커밋 목록과 양방향으로 대조

changelog 는 릴리스의 유일한 사람용 산출물이므로 두 방향이 모두 필요합니다 — 빠진 것(실제 머지됐는데 안 실림)과 없는 것(실재하지 않는 커밋이 실림)입니다. 한쪽만 보면 나머지 절반은 검증되지 않습니다.

$ git show FETCH_HEAD:CHANGELOG.md | awk "/^## \[7\.10\.0\]/{f=1;next} /^## \[7\.9\.1\]/{f=0} f" | grep -oE "commit/[0-9a-f]{7}" | sed "s|commit/||" | sort -u > "$S/cl_shas.txt"
$ git log --format="%h" v7.9.1..origin/main | sort -u > "$S/real_shas.txt"
$ wc -l < "$S/cl_shas.txt" ; wc -l < "$S/real_shas.txt"
      36
      38

두 목록을 comm 으로 양방향 대조합니다:

$ comm -13 "$S/cl_shas.txt" "$S/real_shas.txt" | while read -r s; do git log --format="%h %s" -1 "$s"; done
2062805 chore(memory): normalize frontmatter schema drift (#968)
e0a729a test: isolate CLAUDE_CONFIG_DIR from the test suite (#1004)
$ comm -23 "$S/cl_shas.txt" "$S/real_shas.txt" | wc -l
       0

두 번째 방향이 0 입니다 — changelog 에만 있고 실재하지 않는 커밋은 없습니다. 빈 출력이 아니라 개수로 받은 것은, 빈 출력만으로는 "0건" 과 "명령이 안 돌았음" 이 구분되지 않기 때문입니다. Positive control: 첫 번째 방향이 같은 comm 으로 2건을 찍는 것이 이 프로브가 차이를 실제로 낼 수 있다는 근거입니다.

빠진 2건은 누락이 아니라 설정된 정책입니다 — 정확히 chore 하나와 test 하나이고, release-please 설정이 그 두 타입을 숨기도록 되어 있습니다:

$ git show FETCH_HEAD:release-please-config.json | python3 -c "import json,sys; [print('HIDDEN' if s.get('hidden') else 'SHOWN ', s['type'], '->', s['section']) for s in json.load(sys.stdin)['changelog-sections']]"
SHOWN  feat -> Added
SHOWN  fix -> Fixed
SHOWN  perf -> Changed
SHOWN  refactor -> Changed
SHOWN  docs -> Changed
SHOWN  ci -> Changed
HIDDEN chore -> Changed
HIDDEN test -> Changed
HIDDEN style -> Changed

chore 와 test 가 hidden 이고, 빠진 2건이 정확히 그 두 타입입니다. $S 는 중간 파일을 워크트리 밖에 두려고 담아 둔 경로 변수입니다.

Evidence 3 — 런타임 경로가 그대로임을 종료 코드로 확인

이 확인이 행 4 의 전제입니다. 릴리스 브랜치가 훅·스킬·스크립트를 건드리지 않는다면, main 에서 스위트를 돌리는 것이 곧 릴리스될 내용을 검증하는 것이 됩니다. 반대로 한 줄이라도 다르면 main 실측은 다른 트리에 대한 측정이 되어 무효입니다.

$ git diff --quiet origin/main FETCH_HEAD -- hooks/ skills/ scripts/ tests/ docs/ ; echo "runtime-paths exit=$?"
runtime-paths exit=0

exit=0 이 "차이 없음" 입니다. git diff --quiet 는 답을 종료 코드로 내는 프로브라 빈 출력이 곧 답이고, 그래서 상태를 함께 찍었습니다.

Positive control: 같은 프로브를 이 PR 이 확실히 바꾼 파일에 걸면 exit=1 이 나옵니다 — 없으면 위 exit=0 이 "차이 없음" 인지 "프로브가 아무것도 못 보는 상태" 인지 구분되지 않습니다.

$ git diff --quiet origin/main FETCH_HEAD -- VERSION ; echo "positive control (VERSION) exit=$?"
positive control (VERSION) exit=1

PR 표면의 변경 파일 목록도 같은 이야기를 합니다 — 10개 전부 버전 파일이거나 changelog 입니다:

$ gh pr diff 964 --name-only
.agents/plugins/marketplace.json
.claude-plugin/marketplace.json
.claude-plugin/plugin.json
.cursor-plugin/plugin.json
.opencode/plugin.json
.release-please-manifest.json
CHANGELOG.md
VERSION
gemini-extension.json
plugins/praxis/.codex-plugin/plugin.json
Evidence 4 — 릴리스될 내용에서 CI 진입점 전 단계를 1회 실행

이 릴리스는 19개 PR 을 담고 있고 그중 다수가 저장소 전반의 훅이 공유하는 계층을 건드렸으므로, 부분 pytest 가 아니라 CI 워크플로가 부르는 진입점을 그대로 돌립니다. Evidence 3 이 릴리스 브랜치와 main 의 런타임 경로가 동일함을 보였으므로, main 에서의 실행이 릴리스될 내용에 대한 실행입니다.

$ env -u PRAXIS_ASK_END_STRICT bash scripts/run-tests.sh > "$LOG" 2>&1 ; echo "exit=$?"
exit=1

$ grep -E '^=== (pytest|shell tests|manifest check|hook-token invariant check|memory frontmatter lint|ruff|shellcheck|markdownlint)|^[0-9]+ passed|^All checks|^TEST SUITE|^plugin-manifest|^hook-token-invariant|frontmatter schema violation|^no changed markdown' "$LOG" | sed "s|$HOME|<HOME>|g"
=== pytest ===
813 passed, 1 xfailed, 16943 warnings in 18.19s
=== shell tests ===
=== manifest check ===
plugin-manifest check OK
=== hook-token invariant check ===
hook-token-invariant check OK (7 invariants verified)
=== memory frontmatter lint ===
6 frontmatter schema violation(s) across 52 memory entries in <HOME>/.claude-4/projects/…/memory
=== ruff ===
All checks passed!
=== shellcheck ===
=== markdownlint (advisory) ===
no changed markdown files
TEST SUITE FAILED

=== shell tests === 아래 개별 스위트 수천 줄(전체 5035 줄)은 접혀 있고, 전부 0 실패인지는 따로 확인했습니다 — 걸린 두 줄은 케이스 이름에 failed 가 들어간 것뿐입니다:

$ grep -nE 'failed: [1-9]|[1-9][0-9]* failed|FAIL: [1-9]|FAIL=[1-9]|^FAIL: [a-z]|^FAIL ' "$LOG" | grep -v 'feedback_'
1932:PASS  [FE15 failed browser MCP call is not observation]
1933:PASS  [FE16 failed Edit is not a surface change]

실패한 단계는 memory frontmatter lint 하나뿐이고 Unverified (b) 에서 다룹니다 — grep -v 'feedback_' 는 그 단계가 찍는 파일명을 제외한 것입니다. env -u PRAXIS_ASK_END_STRICT 는 제 셸에 그 변수가 설정돼 있어 붙였습니다(없으면 advisory 모드 케이스 2건이 strict 모드로 돌아 실패하는데, 하네스 인공물이지 코드 결함이 아닙니다). 홈 경로는 $HOME 를 환경변수로 둔 sed 로 치환했고 명령은 실제로 그렇게 실행했습니다. 그 뒤에 남은 프로젝트 슬러그도 사용자명을 담고 있어 … 로 한 번 더 줄였습니다 — 이 한 곳만 붙여넣은 출력에 손을 댄 자리입니다.

813 passed 는 이번 릴리스의 마지막 두 PR(#1014, #1030)이 머지된 뒤의 main 수치입니다. CI 도 이 PR 의 head 에서 test 잡을 pass 로 마쳤습니다 — 위 로컬 실패가 CI 에 없는 memory 디렉터리 때문이라는 것의 대조입니다.

측정 시점: origin/main = df6cc4e. Evidence 3 이 이 트리와 01832ec 의 런타임 경로가 동일함을 보였습니다.

Evidence 5 — 태그·릴리스의 실재와, 배포 표면이 내놓는 버전을 저장소 밖 경로로 확인

머지 전에는 대상 아티팩트가 존재하지 않아 측정 자체가 불가능했던 항목입니다(rev 1 의 Unverified (a)). 머지로 아티팩트가 생겼으므로 순서대로 세 가지를 봅니다.

먼저 릴리스가 실제로 게시됐는지 — draft 나 prerelease 로 남아 있으면 아무도 받지 못합니다:

$ gh release view v7.10.0 --json tagName,createdAt,isDraft,isPrerelease,url --jq "\(.tagName) created=\(.createdAt) draft=\(.isDraft) prerelease=\(.isPrerelease) \(.url)"
v7.10.0 created=2026-08-16T13:20:24Z draft=false prerelease=false https://github.com/devseunggwan/praxis/releases/tag/v7.10.0

다음으로 태그가 가리키는 커밋이 머지된 main 인지 — 태그가 엉뚱한 커밋에 붙으면 릴리스 노트와 실제 코드가 갈립니다:

$ git fetch --tags origin 2>&1 | tail -1; git rev-parse --short v7.10.0; git rev-parse --short origin/main
 * [new tag]         v7.10.0    -> v7.10.0
af17569
af17569

마지막으로 배포 표면이 무엇을 내놓는지 입니다. 로컬 클론은 이미 행 1 이 읽었으므로 같은 것을 두 번 읽는 셈이 됩니다 — 클라이언트가 실제로 가져가는 경로를 GitHub API 로 직접 받아 독립적으로 봅니다:

$ for f in .claude-plugin/marketplace.json .claude-plugin/plugin.json .agents/plugins/marketplace.json gemini-extension.json; do printf "%-40s " "$f"; gh api "repos/devseunggwan/praxis/contents/$f?ref=main" --jq ".content" | base64 -d | grep -oE "[0-9]+\.[0-9]+\.[0-9]+" | sort -u | tr "\n" " "; echo; done
.claude-plugin/marketplace.json          7.10.0 
.claude-plugin/plugin.json               7.10.0 
.agents/plugins/marketplace.json         7.10.0 
gemini-extension.json                    7.10.0 

Positive control: 같은 API 프로브를 이전 태그에 걸면 이전 버전이 나옵니다 — 없으면 위 출력이 "배포 표면이 갱신됐다" 인지 "이 프로브가 늘 같은 값을 찍는다" 인지 구분되지 않습니다.

$ gh api "repos/devseunggwan/praxis/contents/.claude-plugin/plugin.json?ref=v7.9.1" --jq ".content" | base64 -d | grep -oE "[0-9]+\.[0-9]+\.[0-9]+" | sort -u
7.9.1

측정 시점: af17569(머지 커밋). 남은 구멍은 Unverified (a) 입니다 — 여기까지가 "제공된다" 이고, 클라이언트가 받아 기동하는 것은 저장소 밖 관측입니다.

History — rev 1 → rev 2

rev 1 (01832ec) — 머지 전 게시. 행 1~4 로 릴리스에 담길 내용을 검증했고, 배포 아티팩트가 아직 존재하지 않아 태그·릴리스·배포 표면은 Unverified (a) 로 열어 두었습니다.

rev 2 (01832ec) — 코드 변경 없음, 헤딩 SHA 도 그대로입니다. 머지(af17569)로 아티팩트가 생겨 rev 1 의 Unverified (a) 중 측정 가능해진 부분을 행 5 와 Evidence 5 로 승격했습니다 — 릴리스 게시 여부, 태그가 가리키는 커밋, 그리고 클라이언트가 읽는 경로를 GitHub API 로 직접 받은 버전입니다. (a) 는 삭제하지 않고 남은 범위("제공된다"와 "동작한다" 사이)로 좁혔습니다. 행 14 와 Evidence 14 는 그대로입니다.

@devseunggwan
devseunggwan merged commit af17569 into main Aug 16, 2026
10 checks passed
@devseunggwan
devseunggwan deleted the release-please--branches--main branch August 16, 2026 13:20
@devseunggwan

Copy link
Copy Markdown
Owner Author

🤖 Created releases:

🌻

@devseunggwan

Copy link
Copy Markdown
Owner Author

Verification updated — 01832ec rev 2 · 코드 변경 없음. 머지(af17569)로 배포 아티팩트가 생겨, rev 1 에서 "순서상 측정 불가" 로 열어 두었던 Unverified (a) 중 측정 가능해진 부분을 행 5 · Evidence 5 로 승격했습니다 — 릴리스가 draft/prerelease 가 아닌 게시 상태, 태그 v7.10.0 이 머지 커밋 af17569 를 가리킴, 그리고 클라이언트가 읽는 경로를 GitHub API 로 직접 받아 7.10.0 확인(양성 대조: 같은 프로브를 v7.9.1 에 걸면 7.9.1). (a) 는 삭제하지 않고 남은 범위 — 클라이언트가 받아 실제로 기동하는지는 저장소 밖 관측 — 로 좁혔습니다. → #964 (comment)

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment