Skip to content

[codex] Branch 9W4E first-run model setup and packaged readiness - #134

Open
Ian747-tw wants to merge 29 commits into
mainfrom
redesign/branch9w4e-first-run-model-setup-v2
Open

Ian747-tw wants to merge 29 commits into
mainfrom
redesign/branch9w4e-first-run-model-setup-v2

Conversation

@Ian747-tw

@Ian747-tw Ian747-tw commented Aug 26, 2026 •

Copy link
Copy Markdown
Owner

Summary

Implements first-run Commander/Executor model setup on the packaged OpenCode Executor-readiness command.

  • Persists credential-free runtime_model_setup_committed authority append-only and activates it only on the next Runtime construction.
  • Provides code-owned Claude Sonnet 4.5, Gemini 2.5 Flash, and GPT-4.1 mini recipes for Commander and primary Executor.
  • Uses one immutable packaged OpenCode target for readiness and Executor launch.
  • Requires setup authority before Runtime startup; interactive and headless cancellation cannot escape into lifecycle activity.
  • Preserves independent role credentials and authority, one exact primary --model, and disabled capability flags.

Exact Revision

  • Base: 5c114669ebe8b6fd3ddd50a912a6ff821b882edd
  • Head: 47c4305877352102e314966eb4762359b5987087
  • Ancestry: 0 behind / 29 ahead
  • Open, mergeable, non-draft, unmerged

Commit Map

  1. b913b8d331f0e73ed8ca681b99bcd79ce2877020 M9W4E: record first-run model setup boundary
  2. 3898a25749c7abfcb335d8b8106f97c5abd75359 M9W4E: add immutable setup catalog and durable authority
  3. 957721ea30590656a611ace6575f37f78c7c5d5e M9W4E: add packaged Executor readiness integration
  4. 3dfd81ac07b86e9b9555fb7041ab5a0a03b19342 M9W4E: activate setup during next Runtime construction
  5. ebf304470caa9aeb897998f94fabebce29ceb61b M9W4E: add OpenTUI role-model onboarding
  6. 9585062fc9ccdd114add71b7ef11f0dc7ab13d65 M9W4E: prove restart readiness and launch through real user flow
  7. 637eb7719cf21fc0ca9da6e6afda2ece514c7363 M9W4E: harden packaged readiness identity evidence
  8. 4948ca71b088b6daf2107175875dae6562757877 M9W4E: align first-run onboarding evidence
  9. 055ad03d020121256894816d000671cde8153b32 M9W4E: preserve established setup authority flows
  10. 26de3cc0afb0254b3d81737e16761c96a9f4f9c1 M9W4E: preserve historical no-mutation evidence
  11. 98c617fdec6f7b7d6e3abb81bf362945076dccf2 M9W4E: record local validation evidence
  12. 3b869598b54c042a4b22f8a5a1e657cae27acc7c M9W4E: preserve first-run setup navigation
  13. 10ee3ec36b62122828dfbc48c8d5145d7c099ce0 M9W4E: retain restart-required setup ownership
  14. 51f617444c6944d26b6736a6db4e564e83bb3dd3 M9W4E: enforce one OpenCode process authority
  15. addccb6fa161b531bd96bfec4eb2ac5da730e0c5 M9W4E: gate first-run runtime startup
  16. 638449f8ab9fd46204ff6e1c5338513814d395c1 M9W4E: fence interactive startup on setup authority
  17. 4809fe9630c6b9730f9c36886b9bf78ed758a10c M9W4E: fail closed on setup inspection errors
  18. 69e6a082fb083177a1d3462cb91a7bfeea10179f M9W4E: close interactive and project-root startup bypasses
  19. 050a61c1406facb100bd14680acfea440adf424b M9W4E: validate and recover setup startup inspection
  20. e55a5f3da5d3f1221d3ed78c9641ad551704e1af M9W4E: enforce exclusive complete setup authority
  21. 5fd9b705720f97acb688e9ed70f29d4a724cf87f M9W4E: reject inconsistent and unavailable setup views
  22. c679f70811b4afcf996249037909cbc6ecf3c81e M9W4E: serialize setup mutation against startup
  23. aed33149e937fc5438ee85fadd8f195dd2d7d679 M9W4E: reserve startup and setup authority atomically
  24. 4a0f7d15276562295add85e2434470e832dc7488 M9W4E: serialize readiness inspection with startup
  25. c28a6ee91c27fde57d902db09f8e8b48908ef170 M9W4E: bound readiness observer termination
  26. d53ce33cc73c2a1a5b6d435e22070894143ed3c5 M9W4E: latch readiness termination authority
  27. 462f64099dd96252f6c094b3ab15b38a52dc1b9f M9W4E: gate headless resume on model setup
  28. 0b8b6a6aedf2167fd7260cd1c4c8bed988b6c94a M9W4E: enforce packaged Executor startup authority
  29. 47c4305877352102e314966eb4762359b5987087 M9W4E: require model setup before Runtime startup

Exact-Head Validation

$ cd agentcore/runtime && /usr/bin/time -p bun test
1051 pass
0 fail
9741 expect() calls
Ran 1051 tests across 11 files. [143.09s]
real 143.14
user 86.58
sys 32.22

$ cd agentcore/runtime && bun run typecheck
$ tsc --noEmit

$ cd agentcore/tui && /usr/bin/time -p bun test
357 pass
0 fail
4273 expect() calls
Ran 357 tests across 7 files. [6.94s]

$ cd agentcore/tui && bun run typecheck
$ tsc --noEmit

$ uv run pytest tests/integration/cli -q
.......                                                                  [100%]
7 passed in 2.60s

$ uv run pytest tests/e2e_user/scenarios/test_model_setup_executor_readiness_tui.py -q
.                                                                        [100%]
1 passed in 57.70s

Fresh detached worktree at exact head:

$ cd agentcore/upstream && npx -y bun@1.3.13 install --frozen-lockfile
bun install v1.3.13 (bf2e2cec)
4757 packages installed [6.11s]
real 7.00
user 8.46
sys 7.89

$ npx -y bun@1.3.13 install --frozen-lockfile
bun install v1.3.13 (bf2e2cec)
Checked 2420 installs across 2671 packages (no changes) [1478.00ms]
real 8.15
user 7.55
sys 0.90

$ npx -y bun@1.3.13 run --cwd packages/opencode build:nexusloop-readiness
Smoke test passed: 0.0.0--202609030845
real 9.82
user 12.26
sys 1.87

$ npx -y bun@1.3.13 run --cwd packages/opencode test:nexusloop-readiness-package
packaged readiness executable: pass
source-tree dependency resolution: absent
persistent writes: 0
provider/model requests: 0
real 8.62
user 9.24
sys 0.95

$ cd agentcore/runtime && bun install --frozen-lockfile
bun install v1.3.11 (af24e281)
19 packages installed [139.00ms]
real 0.14
user 0.01
sys 0.09

$ cd agentcore/tui && bun install --frozen-lockfile
bun install v1.3.11 (af24e281)
179 packages installed [191.00ms]
real 0.19
user 0.03
sys 0.17

$ uv run pytest tests/e2e_user -q
........................................................................ [ 80%]
..................                                                       [100%]
90 passed in 1805.58s (0:30:05)
real 1822.18
user 1357.82
sys 468.16

Mechanical checks:

$ git rev-list --left-right --count 5c114669ebe8b6fd3ddd50a912a6ff821b882edd...HEAD
0       29
$ git diff --check
$ git status --short
$ frozen-path guard
all FROZEN.lock paths unchanged
$ git diff --quiet 5c114669e...HEAD -- agentcore/upstream
agentcore/upstream unchanged
$ manifest/lockfile guard
manifests and lockfiles unchanged
$ Commander/OpenCode import guard
no production Commander import/read of OpenCode auth/config/provider/catalog authority added

Current-head gates:

The earlier historical run in phases/9W4E/VALIDATION.md is superseded and excluded. The passing historical evidence above is a new detached run on the clean-reviewed exact head. Development failures and invalid invocations remain documented there.

Exhaustive Changed Files

M agentcore/adr/ADR-035-unified-model-profiles-and-role-bindings.md
M agentcore/adr/ADR-036-runtime-model-profile-registry-and-role-readiness.md
A agentcore/adr/ADR-040-first-run-model-setup-and-role-selection.md
M agentcore/runtime/src/authority/command-authority-registry.ts
M agentcore/runtime/src/authority/command-authority-types.ts
M agentcore/runtime/src/events/event-store.ts
M agentcore/runtime/src/index.ts
M agentcore/runtime/src/launch-config.ts
M agentcore/runtime/src/model-configuration/model-profile-runtime-registry-types.ts
A agentcore/runtime/src/model-configuration/model-setup.test.ts
A agentcore/runtime/src/model-configuration/model-setup.ts
A agentcore/runtime/src/model-configuration/opencode-executor-readiness-resolver.test.ts
A agentcore/runtime/src/model-configuration/opencode-executor-readiness-resolver.ts
M agentcore/runtime/src/runtime.test.ts
M agentcore/runtime/src/server.ts
M agentcore/runtime/src/tui/runtime-client.ts
M agentcore/runtime/src/tui/runtime-server-client.ts
M agentcore/tui/src/app.tsx
M agentcore/tui/src/keyboard.ts
M agentcore/tui/src/launch.ts
A agentcore/tui/src/model-setup-view.ts
M agentcore/tui/src/reducer.ts
M agentcore/tui/src/runtime-client-factory.ts
M agentcore/tui/src/runtime-effects.ts
M agentcore/tui/src/runtime-state-merge.ts
M agentcore/tui/src/runtime.ts
M agentcore/tui/src/snapshot.ts
M agentcore/tui/src/state.ts
M agentcore/tui/test/keyboard.test.ts
M agentcore/tui/test/launch.test.ts
M agentcore/tui/test/reducer.test.ts
M agentcore/tui/test/runtime-client-factory.test.ts
M agentcore/tui/test/runtime-effects.test.ts
M agentcore/tui/test/runtime-state-merge.test.ts
M docs/ARCHITECTURE.md
M docs/COMMANDER_PROVIDERS.md
M docs/TUI_UX.md
A phases/9W4E/BASELINE.md
A phases/9W4E/FORBIDDEN.md
A phases/9W4E/IMPLEMENTATION_BRIEF.md
A phases/9W4E/MODEL_SETUP_SURFACE_AUDIT.md
A phases/9W4E/SCOPE_QUESTIONS.md
A phases/9W4E/THREAT_MODEL.md
A phases/9W4E/VALIDATION.md
A phases/9W4E/checklist.md
M tests/e2e_user/sandbox.py
M tests/e2e_user/scenarios/test_commander_executor_review_tui.py
M tests/e2e_user/scenarios/test_context_budget_registry_tui.py
M tests/e2e_user/scenarios/test_context_packet_compiler_tui.py
M tests/e2e_user/scenarios/test_executor_review_proposal_draft_tui.py
A tests/e2e_user/scenarios/test_model_setup_executor_readiness_tui.py
M tests/e2e_user/scenarios/test_opencode_process_smoke_tui.py
M tests/e2e_user/scenarios/test_research_memory_novelty_tui.py

Diff Stat

53 files changed, 5278 insertions(+), 113 deletions(-)

Scope Exclusions

No credential storage/login UI, OpenCode auth.json mutation, source observer, alternate readiness executable, provider/model discovery, dynamic provider loading, fallback, failover, retry, streaming, hot reload, auxiliary OpenCode model selection, external MCP/research, proposal/governance/mutation authority, upstream change, manifest/lockfile change, or frozen-file change was added. resume_supported=false, provider_tool_loop_enabled=false, and external_read_execution_enabled=false remain unchanged.

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

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

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 98c617fdec

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

Comment thread agentcore/tui/src/runtime-state-merge.ts Outdated
@Ian747-tw

Copy link
Copy Markdown
Owner Author

@codex review the exact head 98c617f for Branch 9W4E first-run Commander/Executor model setup and packaged Executor readiness integration. Current-head CI is green; fresh detached Runtime/TUI/CLI/targeted/historical validation is complete (historical 90/0); and the branch preserves restart-only activation, exact packaged-command identity, role isolation, and disabled capability flags. Please report P1/P2/P3 findings or leave a clean exact-head top-level comment.

@Ian747-tw

Copy link
Copy Markdown
Owner Author

@codex review the exact head 3b86959 for Branch 9W4E. The prior startup-navigation P1 is fixed: authoritative missing setup may supersede only the stream-driven boot-to-resume transition, while intervening operator resume selection remains preserved. Direct regression plus the full TUI suite (344 pass, 0 fail) and typecheck pass; both current-head CI checks are green; the prior thread is resolved and outdated. Please report P1/P2/P3 findings or leave a clean exact-head top-level comment.

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

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

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 3b869598b5

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

Comment thread agentcore/tui/src/keyboard.ts
@Ian747-tw

Copy link
Copy Markdown
Owner Author

@codex review the exact head 10ee3ec36d3c402a809bb22f79ee24d7316a468f for Branch 9W4E. The prior Escape P2 is fixed: committed setup with pending_restart=true now owns both Enter and Escape until process restart, while unchanged committed setup can still return to its origin. Direct regression plus focused/full TUI suites (344 pass, 0 fail) and typecheck pass; both current-head CI checks are green; all live threads are resolved. Please report P1/P2/P3 findings or leave a clean exact-head top-level comment.

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

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

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 10ee3ec36b

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

Comment thread agentcore/runtime/src/launch-config.ts Outdated
@Ian747-tw

Copy link
Copy Markdown
Owner Author

@codex review the exact head 51f6174440d4e5937f64338984765275086acf45 for Branch 9W4E. The prior process-authority P2 is fixed: persisted Executor setup now rejects openCodeAdapterFactoryOptions.spawn in addition to direct launch adapter/spawn overrides, preventing readiness and launch from using different process authority. The regression proves rejection before the injected spawn runs; focused Runtime is 596 pass, full Runtime is 1044 pass, typecheck passes; both current-head CI checks are green; all live threads are resolved. Please report P1/P2/P3 findings or leave a clean exact-head top-level comment.

@chatgpt-codex-connector

Copy link
Copy Markdown

Codex Review: Didn't find any major issues. Delightful!

Reviewed commit: 51f617444c

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

@Ian747-tw

Copy link
Copy Markdown
Owner Author

@codex review the exact head addccb6 for Branch 9W4E. The first-run startup P1 is fixed: model setup status is the first no-start mount check, missing or first-activation-pending setup returns before runtime.status and all broad reads, and complete-journal regressions prove zero adapter/process starts, scheduler bootstrap, runtime_started, or runtime_shutdown before commit and fresh reconstruction. Focused historical scenarios inspect the full journal (6 passed); full Runtime is 1044 pass, full TUI is 347 pass, CLI is 7 pass, targeted production setup E2E is 1 pass; both current-head CI checks are green. Please report P1/P2/P3 findings or leave a clean exact-head top-level comment.

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

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

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: addccb6fa1

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

Comment thread agentcore/tui/src/runtime-effects.ts Outdated
@Ian747-tw

Copy link
Copy Markdown
Owner Author

@codex review the exact head 638449f for Branch 9W4E. The prior interactive startup P1 is fixed: ProjectInitialized remains on a noninteractive boot surface while the no-start setup check is pending, keyboard dispatch is disabled during that window, and a required setup screen cannot be displaced by later initialization events. The regression covers pending, required, and clear navigation/input states; focused TUI is 75 pass, full TUI is 348 pass with typecheck green; both current-head CI checks are green; the prior thread is resolved. Please report P1/P2/P3 findings or leave a clean exact-head top-level comment.

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

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

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 638449f8ab

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

Comment thread agentcore/tui/src/runtime-effects.ts Outdated
@Ian747-tw

Copy link
Copy Markdown
Owner Author

@codex review

@Ian747-tw

Copy link
Copy Markdown
Owner Author

@codex review

Please perform a fresh detailed review of exact head c679f70.

The Runtime now rejects model-setup confirmation for the entire owned startup operation, including pre-lock and post-authority-check readiness/adapter awaits. A delayed-start regression proves no setup event can append across startup. Please inspect this lifecycle serialization together with the complete prior 9W4E authority and TUI startup boundary.

Both exact-head CI checks pass, and the complete local Runtime suite passes 1046 tests.

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

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

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: c679f70811

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

Comment thread agentcore/runtime/src/server.ts
@Ian747-tw

Copy link
Copy Markdown
Owner Author

@codex review

Please perform a fresh detailed review of exact head aed3314.

Startup and setup mutation now reserve one shared in-process mutex before either path awaits the filesystem run lock. Setup calls arriving after startup reject immediately; setup calls already waiting when startup begins yield lifecycle ownership to startup without either operation colliding on the run lock. Regressions cover both orderings, zero setup append, successful startup, and shutdown overlap.

Both exact-head CI checks pass and the complete local Runtime suite passes 1047 tests.

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

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

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: aed33149e9

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

Comment thread agentcore/runtime/src/model-configuration/opencode-executor-readiness-resolver.ts Outdated
@Ian747-tw

Copy link
Copy Markdown
Owner Author

@codex review

Please review exact head 4a0f7d15276562295add85e2434470e832dc7488 in detail. This head makes production Executor readiness startup wait for an already-owned pre-start runtime.model_setup_status observation without cancelling it, rejects new observations during the bounded startup transition, and preserves shutdown cancellation/draining. Regressions cover the resolver directly and the supported RuntimeServer command/start race. Local Runtime: 1049 pass, 0 fail; runtime typecheck passed. Exact-head CI runs 33099352796 and 33099357030 are green.

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

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

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 4a0f7d1527

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

@Ian747-tw

Copy link
Copy Markdown
Owner Author

@codex review

Please review exact head c28a6ee91c27fde57d902db09f8e8b48908ef170 in detail. This head retains the startup/inspection lifecycle barrier and now guarantees bounded observation settlement when a child never emits close: timeout or shutdown sends SIGTERM, then SIGKILL after the fixed grace period, settles with the fixed bounded error, and removes active ownership. Regressions cover non-closing children under both startup timeout and shutdown. Local Runtime: 1050 pass, 0 fail; runtime typecheck passed. Exact-head CI runs 33100353718 and 33100359056 are green.

@Ian747-tw

Copy link
Copy Markdown
Owner Author

@codex review

Renewed exact-head review request for c28a6ee91c27fde57d902db09f8e8b48908ef170; the prior request at issue comment 5443065160 was acknowledged but produced no review record after eleven minutes. Please perform a detailed review of this exact SHA. Both exact-head CI runs are green, and local Runtime validation is 1050 pass / 0 fail with typecheck passing.

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

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

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: c28a6ee91c

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

Comment thread agentcore/runtime/src/model-configuration/opencode-executor-readiness-resolver.ts Outdated
@chatgpt-codex-connector

Copy link
Copy Markdown

Codex Review: Didn't find any major issues. You're on a roll.

Reviewed commit: c28a6ee91c

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

@Ian747-tw

Copy link
Copy Markdown
Owner Author

@codex review

Please review exact head d53ce33cc73c2a1a5b6d435e22070894143ed3c5 in detail. Termination now latches the first bounded reason, clears the observation timeout when termination begins, ignores child error events during termination so SIGKILL escalation remains scheduled, and settles/removes ownership after the fixed grace period even if close never arrives. The hostile-child regression covers error-after-SIGTERM and shutdown overlapping the original timeout. Local Runtime: 1050 pass, 0 fail; runtime typecheck passed. Exact-head CI runs 33101973917 and 33101979245 are green.

@chatgpt-codex-connector

Copy link
Copy Markdown

Codex Review: Didn't find any major issues. Keep them coming!

Reviewed commit: d53ce33cc7

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

@Ian747-tw

Copy link
Copy Markdown
Owner Author

@codex review

Renewed detailed review request for exact head d53ce33cc73c2a1a5b6d435e22070894143ed3c5. The prior request at issue comment 5443296214 lost its acknowledgment without producing a review record. Please review this exact SHA; both current-head CI workflows are green and local Runtime validation is 1050 pass / 0 fail with typecheck passing.

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

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

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: d53ce33cc7

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

Comment thread agentcore/tui/src/launch.ts
@Ian747-tw

Copy link
Copy Markdown
Owner Author

@codex review

Please review exact head 462f64099dd96252f6c094b3ab15b38a52dc1b9f in detail. Headless no-start scripts now perform the server-owned model-setup inspection before any explicit resume handling. Missing setup enters the setup screen, so scripted submit plus /opencode-session-plan cannot auto-start Runtime, the adapter, or scheduler; active setup preserves existing no-start behavior. A real RuntimeServer client regression proves created lifecycle, no run lock, zero adapter starts, and zero events. Local TUI: 355 pass, 0 fail; TUI typecheck passed. Exact-head CI runs 33103637306 and 33103643058 are green.

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

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

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 462f64099d

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

Comment thread agentcore/runtime/src/launch-config.ts Outdated
@Ian747-tw

Ian747-tw commented Aug 27, 2026 •

Copy link
Copy Markdown
Owner Author

@codex review exact head 0b8b6a6

@Ian747-tw

Copy link
Copy Markdown
Owner Author

@codex review

Please review exact head 0b8b6a6aedf2167fd7260cd1c4c8bed988b6c94a in detail. Persisted or explicit Executor selection now requires one packaged process target at the production launch factory; direct Runtime adapters, missing/fake process configs, injected readiness resolvers, and alternate launch spawns fail before Runtime construction. Production-path fixtures use that same process executable for readiness, Runtime startup, and tactical launch. Local Runtime: 1050 pass, 0 fail; Runtime typecheck passed. Local TUI: 355 pass, 0 fail; TUI typecheck passed. Exact-head CI runs 33105268952 and 33105274582 are green.

@chatgpt-codex-connector

Copy link
Copy Markdown

Codex Review: Didn't find any major issues. Can't wait for the next one!

Reviewed commit: 0b8b6a6aed

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

@Ian747-tw

Copy link
Copy Markdown
Owner Author

@codex review

Please review exact head 47c4305877352102e314966eb4762359b5987087 in detail. Required first-run setup is now confined to the model-setup screen; Escape cannot leave its mandatory entry stage, and every headless command passes through the same gate. Production Runtime startup now rejects absent committed setup and absent permitted external model authority before adapter startup, lifecycle events, or scheduler bootstrap. Local Runtime: 1051 pass, 0 fail; Runtime typecheck passed. Local TUI: 357 pass, 0 fail; TUI typecheck passed. CLI integration: 7 passed. Targeted real setup E2E: 1 passed.

@chatgpt-codex-connector

chatgpt-codex-connector Bot commented Sep 3, 2026 •

Copy link
Copy Markdown

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review Status Commit Review trigger
📝 Code Review ✅ Completed 2026-09-03T08:44:55.324752Z 47c4305 Manual request
ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

@chatgpt-codex-connector

Copy link
Copy Markdown

Codex Review: Didn't find any major issues. Hooray!

Reviewed commit: 47c4305877

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

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.

1 participant