Skip to content

Fix approval ownership, runtime reconciliation and live views - #739

Merged
Varuas37 merged 26 commits into
mainfrom
codex/691-unblocked-audit
Oct 11, 2026
Merged

Varuas37 merged 26 commits into
mainfrom
codex/691-unblocked-audit

Conversation

@Varuas37

@Varuas37 Varuas37 commented Oct 10, 2026 •

Copy link
Copy Markdown
Contributor

Motivation

Complete the unblocked work from audit tracker #691, following the user's request to pause issues labelled blocked-by-restructure. The earlier combined local branch is preserved; #678, #680, #681, #682, #683, #686 and #696 are deferred until their runtime-architecture slices are ready. This PR does not close the tracker.

#692 and #694 already fixed #679 and #687; #715 covered #684 and superseded closed, unmerged #677. This PR completes #677's remaining mounted panel Messages-list goal as #722 through the current subscription contract.

Solution

  1. Keep SDK approval-change ownership through provider acknowledgment, caller loss and close ([P1] Pending SDK approval changes can block close and lose acknowledgement on caller drop #685; 15cfad98f, 4ef66c8ff).
  2. Accept safely owned shared service directories, reconcile exact installed commands and changed environment, and retain one directory-publication transaction through physical settlement and audit failure ([P1] Existing service directories with 0755 permissions block packaged gateway startup #688–[P2] Reopening the Linux desktop after a shell PATH change preserves the gateway's old environment #690; 9460a266f, 4baa5c5da, 5c949fc22, 68f53229d).
  3. Refuse a retained OAuth refresh secret from another authorization generation before making the token request ([P2] OAuth refresh accepts a retained secret from another authorization generation #711; 78d8a4958).
  4. Correlate metadata targets and newly Created receipts with the original proposal before effects; retain owned mode intent/application through terminal faults; and publish current views through the projection and existing retirement owners ([P1] Metadata replies can falsely downgrade approval mode and leave terminal views stale #712/[P2] Creation metadata reply can attribute audit to an unrequested conversation #726/[P2] Created acknowledgement can replace the admitted creation audit origin #727; adf20f3dd, 43e171c9d). This uses existing current-read and LiveOnly paths without the deferred history API.
  5. Keep one mounted active/archived Messages-list pair, reject late callbacks and one-shot replies, retain each half’s stale evidence, and release late opens on cleanup (Panel Messages list does not follow committed catalogue changes #722; e27825680). Its target budget of two is generated from the existing schema owner; separate list frames are not claimed as an atomic snapshot.
  6. Move ResizeObserver effects onto cancellable frames; retire queued focus settlement with its mounted effect, including both overview-return frames; and measure existing browser readiness before sampling or clicking (Investigate WebKit ResizeObserver page errors in scripted gateway checks #693/[P2] Queued workspace focus settlement can start work after cleanup #730; 69135f63d, 5ed886807, 19d11a3dd, 90eb7bcda). The original layout setup and timing budgets are unchanged. Live mode/list verification uses the shared dev-only fixture classification (2b57adac3, a413aee6b).

The branch was assembled from main 965797a9f, retaining #715 subscriptions, #714/#721 execution leases, #720 read grants and #725/#728/#729 peer admission and client feature wiring. It adds no deferred passive-history API or desktop history renderer changes.

  1. Refuse a bound unavailable view through the existing draft decision before create/send/steer; retain the draft and project the same failure into the existing approval control until valid recovery ([P1] Metadata replies can falsely downgrade approval mode and leave terminal views stale #712; 9c1c672f2). The unchanged headed Chromium/WebKit mode fixture proves both consumers, and a fresh independent delta review reports zero findings.

Current source head is fb8d62152d2c26c9ae6e659b7b02d7f5cadfff38 (native test-gate correction and compact final evidence). Upstream subsequently advanced to 2d2a7a714e053d25b7371d86ff53b8b77952a363 (#732, SDK MCP writer-panic cleanup); a read-only merge preflight is clean. The measured main performance baseline below remains the actual 88f34e4da source, and is not relabelled as the later main. The new CI workflow checkout is independently verified as merge db98edc8718f52b8328880358feea5d1cd8d3678, parents current main2d2a and PRfb8. All 10 checks on that actual current-main merge now pass, including all three supported-platform jobs and the required aggregate.

  1. Keep the native settings fault counter compatible with the declared Rust 1.85 minimum and current CI warnings: use its existing checked decrement with a compare/exchange loop inside the test fixture (31e0ff4ee). Production code and warning policy are unchanged.

  2. Consume the existing complete-list, recorded ghost and resize-frame publications before verification measures them (2040a42b3); refuse an unavailable complete-list observation through the existing CannotRun result (a327700ec). Original product assertions, seeds, selectors, geometry and budgets are unchanged. The readiness review’s P2 is fixed and its fresh second round reports zero findings.

  3. Make the existing two-caller native tests consume both receipt waiters before releasing their held effect (7f9ef1807). The retained publication introduced in 68f53229dd makes its first caller a waiter too; the old nonzero test gate could release before the second caller joined. Three existing gates are corrected, the sole-caller gate stays unchanged, and production code is unchanged. Controlled before/fixed/removal/restored witnesses distinguish the original failure from valid late admission through typed outcomes, directory state and native attempt counts.

Native settings settlement

Previously, losing a caller could restore the old directory while a detached native change still installed the new command. One publication transaction now retains the desired settings through the actual native receipt. A failed audit preserves its confirmed physical result.

sequenceDiagram
    participant Caller
    participant Owner as ClaudePublication transaction
    participant Settings
    participant Native as Native reconciliation
    participant Result as Retained result
    Caller->>Owner: Claim requested directory
    Owner->>Settings: Save desired directory
    Owner->>Native: Reconcile exact owned command
    Caller->>Owner: Drop waiting caller
    Note over Owner: Keep transaction and prior values
    Native-->>Owner: Physical result and audit receipt
    alt Confirmed physical success
        Owner->>Settings: Keep desired directory
    else Native refusal before effects
        Owner->>Settings: Restore known prior values
    else Native effect is unconfirmed
        Owner->>Settings: Restore known local priors and retain failure
        Note over Owner,Native: No daemon rollback is claimed
    end
    Owner->>Result: Settle retained receipt and release ownership
Loading

Messages-list lifetime

The list previously refreshed only on opening, reconnection or local changes. Its mounted owner now receives external catalogue changes through the current subscription API and stops both targets when it leaves.

sequenceDiagram
    participant Messages as Mounted Messages owner
    participant Client as Client subscription gate
    participant Gateway
    participant Other as Other window
    Messages->>Client: Follow active and archived lists
    Client->>Gateway: Subscribe both bounded targets
    Gateway-->>Messages: Initial frame for each half
    Messages->>Messages: Publish roster after both answer
    Other->>Gateway: Create or archive conversation
    Gateway-->>Messages: Replace changed half
    Messages->>Messages: Preserve other half and its stale evidence
    Messages->>Client: Leave and stop both registrations
    Client->>Gateway: Retire targets including late opens
Loading

Focus cleanup

Queued settlement could start a new focus attempt after its subscription had been cleaned up. The mounted effect now owns those outer frames as well as its current attempt. The browser regression replaces the real layout while a movement is queued, delivers the retired callbacks, and checks zero new frame work and a usable replacement composer.

sequenceDiagram
    participant Store
    participant Owner as Mounted focus effect
    participant Frame as Animation-frame queue
    participant Next as Replacement layout
    Store->>Owner: Move focused pane
    Owner->>Frame: Queue owned settlement
    Next->>Owner: Cleanup old layout
    Owner->>Frame: Cancel tracked settlement
    Owner->>Owner: Retire effect and stop current attempt
    Frame-->>Owner: Deliver cancelled callback late
    Owner->>Owner: Refuse work from retired effect
    Next->>Next: Keep current composer usable
Loading

Mode and metadata agreement

A mode-only commit previously reused the old publication revision, and substituted metadata could replace the request or audit origin. The service now keeps its original intent and observed application until a correlated terminal result. Unknown settlement retires the attachment through its existing owner; retained physical uncertainty refuses a clean current replacement. Explicit pending recovery evidence remains marked and read-only.

sequenceDiagram
    participant Service as ConversationService
    participant Metadata
    participant Backend
    participant Audit
    participant Projection
    Service->>Metadata: Begin original immutable mode intent
    Metadata-->>Service: Pending evidence
    Service->>Service: Validate evidence before effects
    Service->>Backend: Apply original requested mode
    Backend-->>Service: Observed application
    Service->>Audit: Record original intent and observed effect
    Service->>Metadata: Settle original request with contained faults
    alt Confirmed matching terminal result
        Service->>Projection: Replace selection through revision owner
    else Terminal result cannot be confirmed
        Service->>Service: Retire through existing attachment owner
        Note over Service,Projection: Retained uncertainty refuses clean current publication
    end
Loading

Loaded and acknowledged metadata must name the requested conversation before access, audit or provider effects. A newly Created receipt must also equal the complete admitted proposal; historical Existing evidence remains legitimate.

sequenceDiagram
    participant Service as ConversationService
    participant Metadata
    participant Audit
    Service->>Metadata: Create admitted immutable proposal
    Metadata-->>Service: Created or historical Existing
    Service->>Service: Validate target and access
    alt Created contradicts original proposal
        Service->>Service: Refuse typed Metadata before effects
    else Correlated Created or legitimate Existing
        Service->>Audit: Reconcile validated original evidence
    end
Loading

Validation and remaining merge blockers

Draft, blocked from merge by the 50 ms overview-answer performance gate. All 10 current-head CI checks pass, including Linux/macOS/Windows and the required aggregate. Full browser verification passes 990/990 at actual checked 05bc3c6c3, with unchanged browser executable source at pushed fb8d62152. The checked production cut is 9c1c672f2f20e39f63ff9745a32d6467259d328d. The initial evidence-only commit preserves its production/verification bytes. Later commits change only the native cfg(test) fault-counter fixture, three verification readiness consumers, three native test sampling gates, their existing owning tables and evidence. Committed evidence and exact source identities separate each run and review head.

  • Full frontend passes 3,998 tests / 307 files, plus formatting, lint/rules, scripts, desktop helpers, architecture, protocol generators, client documentation, client/root types and production build at this source.
  • All twelve Rust verification stages pass at bd8ef0763: exact six-package CI test selection, all-targets Clippy with warnings denied, client-without-CLI/server feature checks, formatting/runtime dependencies, SDK/strict documentation, desktop, all-feature storage, release host and fresh gateway. Rust/native/SDK/OAuth/schema regions are byte-identical at the final consumer cut; reuse is recorded precisely. At the latest pushed head, the native cfg(test) fault fixture and existing three test gates differ from that regional Rust result; their fresh complete host verification is recorded below.
  • Fresh desktop and native/SDK/OAuth reviews report zero findings at 4fea9814, with their assigned regions unchanged. API round 1's panic-payload destructor escape was fixed through one terminal owner at 43e171c9d; its eight public cases and three guard-removal probes hold. Fresh API round 2 reports zero findings at bd8ef0763. The final five-file consumer delta has a separate fresh zero-finding review at 9c1c672f2 (94 scoped tests and two independent ordering probes).
  • Unchanged mode-publication check passes in headed bundled Chromium and WebKit, dev mode: both consumer rows plus console row hold in one run. Ask → Auto → Ask revisions are [1,2,2,3], transcript messages 0; unavailable dispatch delta 0, view-unavailable decline, retained draft and disabled controls; confirmed recovery restores Auto. Both added source rules independently fail on removal and pass after restoration.

Unavailable view retains draft and disables approval in Chromium
Unavailable view retains draft and disables approval in WebKit

Fresh mandatory scripted on final production source 9c1c672f2: Verdict: pass, 37/37 — MCP Apps 16/16, gateway-window 13/13, scripted scenarios 8/8, Chromium and WebKit/default Claude/dev mode; 49.4 seconds. The earlier bd8ef0763 checkpoint remains separately committed. The fresh delegate mode run also passes both engine consumer rows (2/2); it is separate from the author run's two consumer rows plus console row, and these are not additive repeated-trial counts.

Check Chromium WebKit
mcp-apps-gateway pass pass
gateway-window pass pass
scripted-scenarios pass pass

Relevant recorded line: console chromium: requestfailed: http://127.0.0.1:58906/browser/check net::ERR_ABORTED (aborted after a 204 response, #485) (harmless) — the known harmless abort after a 204 response (#485). Raw gateway/ACP/session-capability diagnostics are kept private. The checked source inventory excludes evidence-only files and matches the final production cut; Initial GitHub head 5c1c2257c630a3f5d5dc83ba422c882ecc598fbd added only committed evidence. The later readiness changes do not change the scripted-run executable/helper/fixture regions; this 37/37 result remains attributed to its exact 9c1c672f2 source.

Measured performance is failed, not waived: production/headed bundled Chromium, three runs, 4× throttle, calibrated 120 ms frame measured 120.4 ms (ratio 3.2), 32/35 held at bd8ef0763. Against 50 ms: columns drag-drop max/median 73.6/41.9 ms (one slow frame); columns overview-answer 78.6/74.9 ms (three); sidebar overview-answer 76.7/76.6 ms (three). Answer LoAFs report 55–63 ms script and 37–40 ms forced layout through React’s delegated root keydown dispatcher; drag's LoAF has no script attribution. No source cause or outside-scope disposition is inferred. The measured main88 baseline and failed production9c1 performance are recorded below. Final whole verification passes 990/990 with the corrected readiness consumers; separate retained-app checks passed 4/4 on identical executable source. Deferred cursor-owned #710 is not imported or used to excuse these failures.

Quiet actual-main baseline at 88f34e4dabdeca46486dfbd5294e0a2a00a0a5fa is now measured: production/headed bundled Chromium, five runs, 4× throttle, calibration 120 ms → 120.5 ms (ratio 4.12). Overview-answer fails 5/5 in both layouts: columns max/median 87.9/76.5 ms, sidebar 75.6/73.8 ms; all ten slow frames identify React’s delegated root keydown dispatcher with script 55–68 ms / forced layout 36–45 ms. Drag passes 5/5: columns 42.6/41.8 ms, sidebar 44.4/40.4 ms. This establishes an existing answer-budget failure at the measured main88 source; it neither identifies a source cause nor waives the final gate or equates it with #710. The final-cut full performance run has finished at exact production source 9c1c672f2: 33/35 held, failed. Production/headed bundled Chromium, three runs, 4× throttle, calibration 120 ms → 120.5 ms (ratio 2.94). Only overview-answer fails 3/3 in each layout: columns max/median 76.8/73.3 ms, sidebar 72.7/65.5 ms. All other 31 interaction rows, including both drag rows, pass; the earlier bd8 drag outlier remains a failed historical result, not a claimed fixed defect. Six slow answer LoAFs report script 56–58 ms / forced layout 37–39 ms through React’s delegated dispatcher. The 50 ms gate remains failed.

The complete frozen functional run on production 9c1c672f2 has now ended failed, 20/23 groups held (51m55s), dev mode, Chromium/WebKit, all default layouts/sizes, no quick or omitted functional selection. Drag held 171/172 with a WebKit/sidebar/1000×700 compact-card setup error; columns held 81/84 with three Chromium rail-fold setup errors; command-order held 0/8, missing the expected named overview rows in both engines/layouts. The first failures remain unchanged. Live mode and mounted Messages-list checks pass inside this complete run; widgets, subagents, MCP Apps and production message-sync pass (14/14 for the latter). Its four retained-app dev-only rows are run separately and are not counted as passed here. Separate retained-app checks subsequently passed 4/4 (dev, 7.9 s). Focused probes proved the three verification readiness defects: command rows were read before complete-list publication, compact-card DOM visibility preceded the recorder’s current frame, and rail inputs were sampled before resize publication. The tracked corrections consume existing publications; their original assertions still fail on individual source removals. The initial whole-run failures remain failures.

Read-only tracing identifies the real request-row → overview answer/focus → store command → scenario source → settling/reflow path, with six named source files byte-identical between actual main88 and final production9c1. These are ranked profiling suspects, not causal proof. Historical #642 overview-answer measurements had passed, so neither its older exceptions nor deferred #669/#710 are used to excuse this new measured failure. No deferred performance source is patched. The failed initial whole run is recorded above. The final whole run at exact 05bc3c6c3 passes 23/23 groups, 990/990, exit 0 in 3,112.5 seconds (51m52.5s): Chromium and WebKit, every default layout/size, no quick selection. Drag 172/172, rail/layout 84/84 and command-order 8/8 include the previously failing observations. Internal production message-sync passes 14/14 with three delivery runs per engine/layout: maximum measured delivery frame 34 ms, zero over 50 ms. Chromium is calibrated at 4× throttle; WebKit is unthrottled. The four retained-app dev-only cases remain excluded from that production fixture and separately pass 4/4 on byte-identical executable source. All owned browser/dev/gateway/caffeinate processes ended; this PR stays blocked from merge even if CI passes until the required frame-budget gate is properly resolved.

CodeRabbit explicitly skipped this draft. A fresh independent evidence-only review at pushed 5c1c2257c630a3f5d5dc83ba422c882ecc598fbd reports no new actionable findings, independently verifying all 25 evidence-only additions and exact unchanged source regions. It preserves prior review heads rather than claiming a new whole-code approval. One historical focus probe's preserved-mtime wording is qualified: that probe is not fresh-mtime proof; independent hashes and later exact-source checks establish restored source separately.

The first combined Rust run's generated limits-document mismatch at adf20f3dd remains historical failure evidence; the owner meaning was corrected and regenerated at bd8ef0763, and the exact generator test/full Rust rerun pass. Historical whole-batch reports apply only to their original source.

Initial supported-platform CI at 5c1c2257c failed on all three systems because stable Clippy denied deprecated AtomicUsize::fetch_update in the native settings test fixture; the required aggregate failed too. Its macOS SDK/server suites had passed (1,099 SDK and 2,220 server tests). The correction in 31e0ff4ee preserves the existing checked decrement and exactly-one-fault semantics with Rust-1.85-compatible atomics. Fresh full host verification passes: 705 host tests (one existing ignored), six signed-manifest and nine release-key tests, 279 tooling tests, formatting and all-target Clippy with warnings denied. A source-asserted counter removal fails the public rollback witness; restored fixtures pass 8/8. Independent source review reports zero findings. Native CI correction evidence.

The four-file readiness delta passes 28 affected browser rows (command 8/8, compact drag 8/8, rail 12/12), with 164 shared helper tests, lint and format checks. Individual asserted removals fail their canonical checks and restored source passes. Its round-1 P2—ignoring the complete-list observer's false result—was corrected; the fresh independent round-2 review finds no unresolved priority and independently proves evaluator failure is refused while valid publications remain accepted. Readiness evidence and exact reviewed heads. Targeted checks do not replace the final whole run.

Previous-head CI at 05bc3c6c3 completed: eight jobs passed, including Linux/macOS and SDK coverage; Windows failed one native concurrency witness (427 passed, one failed), and the required aggregate failed. The failing assertion is second.join().unwrap().is_err() in an_identical_in_flight_directory_change_waits_for_that_attempt, src-tauri/tests/gateway/application.rs:2467. The former atomic deprecation is no longer the failure. The waiter gate correction is now pushed at fb8d62152; its controlled public witness reproduces the old assertion failure after compilation, passes with the fix, fails after actual sole-gate removal and passes after fresh restoration. The same scheduling scaffold preserves typed Gateway failures, authoritative host directory restoration and one native attempt under the fixed gate; the broken gate admits a legitimate later Ok request with two attempts. No flaky-test waiver is used. Fresh complete combined host verification at clean source 7f9ef1807 passes 705 tests/one existing ignore, 279 tooling tests, six signed-manifest and nine release-key tests, format/all-target Clippy-Dwarnings in28.2s; fresh independent review is zero. The latest evidence commit changes no executable source. Current native gate correction and review. Actual final whole990/990 report and identity.

Fresh current-head CI at fb8d62152 is PASS: all 10 owned checks, including Linux, macOS, Windows, frontend, gateway contract, release desktop compilation, SDK domain coverage, workflow validation, classifier and the required aggregate. Actual run38078915630 tests merge db98edc8718f52b8328880358feea5d1cd8d3678 with current main2d2a. Backend Windows logs explicitly show the formerly failing identical-caller test passing and 428/428 native tests, plus six manifest and nine release-key tests. The Windows log is truncated; check conclusions and the untruncated required aggregate independently confirm success. Existing platform-specific Unix-only omissions retain their real platform limitations and are not counted as newly passed cases. These actual CI results replace the pending status; local checks were not used as a substitute. Results will be read from this head's backend checks, not inferred from local checks. CodeRabbit continues to decline the draft, which is not an automated approval. The independent05bc assembly replacement report retains its exact source. The new two-file gate correction has a fresh independent review at7f9; a fresh independent currentfb8 assembly/evidence replacement review reports zero unresolved findings at every priority in its stated scope. Completed production reviews retain their actual regions and heads. The draft cannot merge while performance or any other required gate fails.

Verification limits: local macOS injected-adapter and Playwright Chromium/WebKit checks do not establish actual Linux service installation, native WKWebView, signed packaged application, live providers or an external peer deployment. Supported-platform CI is reported separately.

Refs #691
Closes #685
Closes #688
Closes #689
Closes #690
Closes #693
Closes #711
Closes #712
Closes #722
Closes #726
Closes #727
Closes #730

…loss

Refs #685. The direct approval response wait introduced in 4e54bf9 held scheduler ownership in the caller. Keep it on an owned task and run backend construction and response polling through the existing lifecycle control owner. Preserve verified generation mode, fence uncertainty, and extend caller-waker coverage.
Refs #685. Keep Receiver and Notify names in the owning test import block under the canonical Rust import rule.
Refs #688, #689, #690. Retain safe shared service directory modes, validate installed canonical configuration, replace changed commands with fresh generations, and serialize durable settings rollback with Claude publication.
Keep Linux native selection and macOS adapter test compilation paired with their dependencies without enabling this desktop adapter on other Unix targets.
Save durable settings under publication ownership before changing live configuration. Carry callback-owned prior values through failed acknowledgement, roll them back before native dispatch, and reject settings success that never applied the change. Refs #689.
…tion

Refs #711. Check restored and held secret generations before expiry or rejected-bearer refresh effects. Preserve mismatched secrets and retain matching-generation refresh behavior.
Keep pane, widget and nested overview settlement under the mounted focus effect. Cleanup cancels queued handles and prevents late delivery from starting a new target search.

Add public hook lifetime regressions and a Chromium/WebKit layout replacement check while preserving existing focus ordering.

Corrects the unowned pane settlement from 1573ff6 and widget path from 00f9180, retained by overview staging in c13a9ed.
Messages must stay current while another caller changes the catalogue.
The history slice owns one active/archived pair and rejects old callbacks
when commands refresh it or the surface leaves. Publish the two-target
bound from the existing product schema and retain explicit stale facts.

The original snapshot-only path came from 9c90048 (Add the Messages
list, conversation archive and permanent delete). This uses the current
replay-to-live subscriptions rather than restoring the old polling stack.

Refs #722
Record the actual main-based source identity, fixed Chromium and WebKit results, and source-asserted UI and generated-capacity reverts for issue #722.
Metadata acknowledgements must describe the admitted target and origin
before audit or provider effects. Mode publication must retain the
original intent and observed application until a correlated terminal
receipt, and retain stopping ownership until physical cleanup confirms.

Regressions originate in 4e54bf9 (Implement per-conversation model and
approval modes) and cb9121b (Durable command receipts and turn-specific
Stop (#537)); their replies previously replaced admitted evidence.

Keep the correction on the existing current read and LiveOnly paths.
No paused history, paging, desktop or SDK work is included.

Fixes #712
Fixes #726
Fixes #727
A repository panic payload can panic again when destroyed. Dropping it
inside terminal settlement escapes the readback/retirement owner and
leaves a clean current mode inconsistent with durable and provider state.
Use the existing payload-retention idiom through one terminal helper.

Regression introduced by adf20f3 (conversation: preserve mode and
metadata owners). The public matrix covers ordinary and destructor-
panicking payloads at both invocation and polling boundaries.

Fixes #712
…very

Consume the existing typed failed-read fact before draft admission and in the current approval row. Preserve bound draft text, attachments and view identity while refusing create/send/steer; valid follow publication restores legitimate input. Refs #712.
Keep exact source and review identities with current consumer screenshots. Record failed frame budgets and pending browser and current-main gates explicitly before starting the grouped draft PR. Refs #691.
@coderabbitai

coderabbitai Bot commented Oct 10, 2026 •

Copy link
Copy Markdown

Review in Change Stack →

Note

Currently processing new changes in this PR. This may take a few minutes, please wait...

⚙️ Run configuration
  • Configuration used: defaults
  • Review profile: CHILL
  • Plan: Advanced
  • Run ID: 681748a2-6e99-4999-a0dc-fcc0af7b4cf0

📥 Commits

Reviewing files that changed from the base of the PR and between 88f34e4 and fb8d621.


⛔ Files ignored due to path filters (12)
  • packages/nessa-client/src/generated/product.ts is excluded by !**/generated/**
  • verification/desktop/evidence/panel-list-follow/unblocked-main/panel-list-follow-live-chromium.png is excluded by !**/*.png
  • verification/desktop/evidence/panel-list-follow/unblocked-main/panel-list-follow-live-webkit.png is excluded by !**/*.png
  • verification/desktop/evidence/panel-list-follow/unblocked-main/panel-list-follow-stale-chromium.png is excluded by !**/*.png
  • verification/desktop/evidence/panel-list-follow/unblocked-main/panel-list-follow-stale-webkit.png is excluded by !**/*.png
  • verification/desktop/evidence/unblocked-audit/mode-publication/unavailable-chromium.png is excluded by !**/*.png
  • verification/desktop/evidence/unblocked-audit/mode-publication/unavailable-webkit.png is excluded by !**/*.png
  • verification/desktop/evidence/unblocked-audit/native-publication-waiters/before-relevant.log is excluded by !**/*.log
  • verification/desktop/evidence/unblocked-audit/native-publication-waiters/combined-host-relevant.log is excluded by !**/*.log
  • verification/desktop/evidence/unblocked-audit/native-publication-waiters/fixed-relevant.log is excluded by !**/*.log
  • verification/desktop/evidence/unblocked-audit/native-publication-waiters/removed-relevant.log is excluded by !**/*.log
  • verification/desktop/evidence/unblocked-audit/native-publication-waiters/restored-relevant.log is excluded by !**/*.log

📒 Files selected for processing (171)
  • crates/nessa-local-storage/src/lib.rs
  • crates/nessa-local-storage/src/retained_directory.rs
  • crates/nessa-local-storage/src/unix.rs
  • crates/nessa-local-storage/src/unix/retained_directory.rs
  • crates/nessa-local-storage/tests/retained_directory.rs
  • crates/nessa-protocol/src/conversation/projection.rs
  • crates/nessa-protocol/src/conversation/view.rs
  • crates/nessa-protocol/src/product/generated.rs
  • crates/nessa-protocol/tests/conversation/projection.rs
  • crates/nessa-sdk/docs/agent_execution/agent.md
  • crates/nessa-sdk/docs/agent_execution/lifecycle.md
  • crates/nessa-sdk/src/application/agent_execution/agents/agent.rs
  • crates/nessa-sdk/tests/application/agent_execution/agents/approval.rs
  • crates/nessa-sdk/tests/application/agent_execution/agents/review_regressions/caller_wakers.rs
  • crates/nessa-server/src/conversation/application/environment.rs
  • crates/nessa-server/src/conversation/application/metadata_target.rs
  • crates/nessa-server/src/conversation/application/mod.rs
  • crates/nessa-server/src/conversation/application/passive_read.rs
  • crates/nessa-server/src/conversation/application/ports.rs
  • crates/nessa-server/src/conversation/application/read_grants.rs
  • crates/nessa-server/src/conversation/application/service.rs
  • crates/nessa-server/src/conversation/application/service/app_calls.rs
  • crates/nessa-server/src/conversation/application/service/creation.rs
  • crates/nessa-server/src/conversation/application/service/mutation.rs
  • crates/nessa-server/src/limits.rs
  • crates/nessa-server/src/mcp_authorization/application/owner.rs
  • crates/nessa-server/tests/conversation/approval_mode.rs
  • crates/nessa-server/tests/conversation/creation_metadata.rs
  • crates/nessa-server/tests/conversation/desktop_stop.rs
  • crates/nessa-server/tests/conversation/passive_read.rs
  • crates/nessa-server/tests/conversation/support.rs
  • crates/nessa-server/tests/mcp_authorization/mod.rs
  • crates/nessa-server/tests/product/socket/subscriptions.rs
  • docs/ARCHITECTURE.md
  • docs/adr/688-shared-service-directory-permissions.md
  • docs/adr/689-linux-service-definition-replacement.md
  • docs/adr/README.md
  • docs/adr/done/231-model-and-approval-per-conversation.md
  • docs/adr/todo/392-remote-mcp-servers.md
  • docs/codebase-structure.md
  • docs/design/record-subscriptions.md
  • docs/design/ui-workspace-load.md
  • docs/limits.md
  • docs/state/services/desktop/workspace/fold-reveal-and-resize-side-columns.md
  • protocol/product/v1.json
  • src-tauri/Cargo.toml
  • src-tauri/src/gateway/application/mod.rs
  • src-tauri/src/gateway/application/ports.rs
  • src-tauri/src/gateway/application/service.rs
  • src-tauri/src/gateway/domain/value_objects/mod.rs
  • src-tauri/src/gateway/infrastructure/commands.rs
  • src-tauri/src/gateway/infrastructure/linux/reconciliation.rs
  • src-tauri/src/gateway/infrastructure/linux/staging.rs
  • src-tauri/src/gateway/infrastructure/linux/unit.rs
  • src-tauri/src/gateway/infrastructure/macos.rs
  • src-tauri/src/gateway/infrastructure/mod.rs
  • src-tauri/tests/gateway/application.rs
  • src-tauri/tests/gateway/application/claude_publication.rs
  • src/conversation/adapters/gateway/effects.ts
  • src/conversation/adapters/gateway/list-follow.test.ts
  • src/conversation/adapters/scenario/effects.ts
  • src/conversation/adapters/store/history-follow.test.ts
  • src/conversation/adapters/store/history.ts
  • src/conversation/adapters/store/read-failure.test.ts
  • src/conversation/application/ports.ts
  • src/conversation/application/usecases/attachments.test.ts
  • src/conversation/application/usecases/send-draft.ts
  • src/conversation/index.ts
  • src/conversation/ui/conversation-list.test.ts
  • src/conversation/ui/conversation-list.tsx
  • src/desktop/ui/column-header.test.tsx
  • src/desktop/ui/column-header.tsx
  • src/desktop/workspace/adapters/dom/focus.test.tsx
  • src/desktop/workspace/adapters/dom/focus.ts
  • src/desktop/workspace/adapters/dom/width.test.tsx
  • src/desktop/workspace/adapters/dom/width.ts
  • src/desktop/workspace/ui/overview/overview.test.tsx
  • src/panel/ui/app.tsx
  • verification/desktop/CHECKLIST.md
  • verification/desktop/evidence/panel-list-follow/unblocked-main/browser-results.json
  • verification/desktop/evidence/panel-list-follow/unblocked-main/capacity-revert-results.json
  • verification/desktop/evidence/panel-list-follow/unblocked-main/capacity-revert.diff
  • verification/desktop/evidence/panel-list-follow/unblocked-main/main-ui-revert.diff
  • verification/desktop/evidence/panel-list-follow/unblocked-main/main-ui-revert.json
  • verification/desktop/evidence/panel-list-follow/unblocked-main/main-ui-source.json
  • verification/desktop/evidence/panel-list-follow/unblocked-main/pr-summary.md
  • verification/desktop/evidence/panel-list-follow/unblocked-main/rule-revert-results.json
  • verification/desktop/evidence/panel-list-follow/unblocked-main/source-identity.json
  • verification/desktop/evidence/unblocked-audit/README.md
  • verification/desktop/evidence/unblocked-audit/api-author-source.json
  • verification/desktop/evidence/unblocked-audit/api-author.md
  • verification/desktop/evidence/unblocked-audit/api-review-round1.md
  • verification/desktop/evidence/unblocked-audit/api-review-round2.md
  • verification/desktop/evidence/unblocked-audit/api-rule-revert-results.json
  • verification/desktop/evidence/unblocked-audit/assembled-source-identity.json
  • verification/desktop/evidence/unblocked-audit/browser-readiness/P2-guard-removal.json
  • verification/desktop/evidence/unblocked-audit/browser-readiness/author-report.md
  • verification/desktop/evidence/unblocked-audit/browser-readiness/command-metadata.json
  • verification/desktop/evidence/unblocked-audit/browser-readiness/command-result.json
  • verification/desktop/evidence/unblocked-audit/browser-readiness/drag-metadata.json
  • verification/desktop/evidence/unblocked-audit/browser-readiness/drag-result.json
  • verification/desktop/evidence/unblocked-audit/browser-readiness/independent-healthy.json
  • verification/desktop/evidence/unblocked-audit/browser-readiness/independent-negative.json
  • verification/desktop/evidence/unblocked-audit/browser-readiness/rail-metadata.json
  • verification/desktop/evidence/unblocked-audit/browser-readiness/rail-result.json
  • verification/desktop/evidence/unblocked-audit/browser-readiness/round1-review.md
  • verification/desktop/evidence/unblocked-audit/browser-readiness/round2-identity.json
  • verification/desktop/evidence/unblocked-audit/browser-readiness/round2-review.md
  • verification/desktop/evidence/unblocked-audit/browser-readiness/source-manifest.json
  • verification/desktop/evidence/unblocked-audit/consumer-author.md
  • verification/desktop/evidence/unblocked-audit/consumer-mode-browser.json
  • verification/desktop/evidence/unblocked-audit/consumer-review.md
  • verification/desktop/evidence/unblocked-audit/consumer-source.json
  • verification/desktop/evidence/unblocked-audit/desktop-review-source.json
  • verification/desktop/evidence/unblocked-audit/desktop-review.md
  • verification/desktop/evidence/unblocked-audit/final-browser-05bc/README.md
  • verification/desktop/evidence/unblocked-audit/final-browser-05bc/ci-merge-identity.json
  • verification/desktop/evidence/unblocked-audit/final-browser-05bc/cleanup.json
  • verification/desktop/evidence/unblocked-audit/final-browser-05bc/image-readback.md
  • verification/desktop/evidence/unblocked-audit/final-browser-05bc/report.md
  • verification/desktop/evidence/unblocked-audit/final-browser-05bc/source-identity.json
  • verification/desktop/evidence/unblocked-audit/final-browser-05bc/summary.json
  • verification/desktop/evidence/unblocked-audit/final-browser-05bc/windows-diagram-readback.md
  • verification/desktop/evidence/unblocked-audit/focus-author.md
  • verification/desktop/evidence/unblocked-audit/frontend-current-results.json
  • verification/desktop/evidence/unblocked-audit/frontend-results.json
  • verification/desktop/evidence/unblocked-audit/native-ci-compatibility/independent-review.md
  • verification/desktop/evidence/unblocked-audit/native-ci-compatibility/report.md
  • verification/desktop/evidence/unblocked-audit/native-ci-compatibility/restoration.json
  • verification/desktop/evidence/unblocked-audit/native-ci-compatibility/results.json
  • verification/desktop/evidence/unblocked-audit/native-ci-compatibility/revert-results.json
  • verification/desktop/evidence/unblocked-audit/native-ci-compatibility/source-identity.json
  • verification/desktop/evidence/unblocked-audit/native-publication-waiters/README.md
  • verification/desktop/evidence/unblocked-audit/native-publication-waiters/author-report.md
  • verification/desktop/evidence/unblocked-audit/native-publication-waiters/before-source.json
  • verification/desktop/evidence/unblocked-audit/native-publication-waiters/combined-host-results.json
  • verification/desktop/evidence/unblocked-audit/native-publication-waiters/controlled-scheduling.patch
  • verification/desktop/evidence/unblocked-audit/native-publication-waiters/final-restoration.json
  • verification/desktop/evidence/unblocked-audit/native-publication-waiters/final-source.json
  • verification/desktop/evidence/unblocked-audit/native-publication-waiters/fixed-source.json
  • verification/desktop/evidence/unblocked-audit/native-publication-waiters/independent-review.md
  • verification/desktop/evidence/unblocked-audit/native-publication-waiters/independent-source.json
  • verification/desktop/evidence/unblocked-audit/native-publication-waiters/removal-source.json
  • verification/desktop/evidence/unblocked-audit/native-publication-waiters/restored-schedule-source.json
  • verification/desktop/evidence/unblocked-audit/native-publication-waiters/semantic-results.json
  • verification/desktop/evidence/unblocked-audit/native-review.md
  • verification/desktop/evidence/unblocked-audit/native-sdk-results.json
  • verification/desktop/evidence/unblocked-audit/native-source-identity.json
  • verification/desktop/evidence/unblocked-audit/panic-payload-correction.md
  • verification/desktop/evidence/unblocked-audit/performance-bd8-failed.txt
  • verification/desktop/evidence/unblocked-audit/rust-results.json
  • verification/desktop/evidence/unblocked-audit/scripted-bd8.md
  • verification/desktop/fixtures/mode-publication/index.html
  • verification/desktop/fixtures/mode-publication/main.tsx
  • verification/desktop/fixtures/panel-list-follow/index.html
  • verification/desktop/fixtures/panel-list-follow/main.tsx
  • verification/desktop/page-lines.md
  • verification/desktop/scripts/columns.mjs
  • verification/desktop/scripts/command-order.mjs
  • verification/desktop/scripts/drag.mjs
  • verification/desktop/scripts/focus.mjs
  • verification/desktop/scripts/lib/browser.mjs
  • verification/desktop/scripts/lib/cli.mjs
  • verification/desktop/scripts/lib/cli.test.mjs
  • verification/desktop/scripts/lib/selectors.mjs
  • verification/desktop/scripts/mcp-apps-gateway.mjs
  • verification/desktop/scripts/mode-publication.mjs
  • verification/desktop/scripts/panel-list-follow.mjs
  • verification/desktop/scripts/responsive.mjs
  • verification/desktop/scripts/run-all.mjs
  • verification/desktop/scripts/shared-controls.mjs

 ____________________________________________________________________
< Never go full rewrite. You don't buy that? Ask Netscape Navigator. >
 --------------------------------------------------------------------
  \
   \   \
        \ /\
        ( )
      .( o ).
✨ Finishing Touches
📝 Generate docstrings
  • Commit to this branch
  • Create a new PR

🧪 Generate unit tests (beta)
  • Commit to this branch
  • Create a new PR

  • Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

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

@Varuas37

Copy link
Copy Markdown
Contributor Author

Current-source evidence for production cut 9c1c672f2f20e39f63ff9745a32d6467259d328d, preserved by the evidence-only pushed head 5c1c2257c630a3f5d5dc83ba422c882ecc598fbd:

ok    confirmed-mode-consumers [chromium]
ok    confirmed-mode-consumers [webkit]
mode-publication: 2/2 held
Verdict: pass

Fresh mandatory scripted: 37/37, both engines/default Claude/dev mode — MCP Apps 16/16, gateway-window 13/13, scripted scenarios 8/8. Source revisions [1,2,2,3], transcript messages 0; unavailable dispatch delta 0, view-unavailable decline, retained draft, disabled approval, valid Auto recovery. These are separate runs, not additive repeated-trial counts.

Chromium unavailable view retains draft
WebKit unavailable view retains draft

Committed reports, identities and failed performance table. Full current frontend passes 3,998/307; twelve Rust stages and zero-finding regional reviews retain their actual checked heads, with precise unchanged source boundaries. The fresh final consumer review reports zero findings. All five sequence diagrams have been verified in the real GitHub renderer.

Draft merge blockers remain: measured performance failed 32/35 at bd8; current-main baseline, final-cut performance/whole functional/retained-app checks and current-main merge-result supported-platform CI are pending. CodeRabbit's success status explicitly means draft review skipped, not reviewed. Seven labelled paused issues remain excluded; tracker #691 stays open. No budget or failing assertion is waived.

@Varuas37

Copy link
Copy Markdown
Contributor Author

CodeRabbit declined review because this PR is draft; its green status is not an external approval. The fresh independent current-head artifact review below and the explicitly source-qualified regional reviews are the local replacement for their actual scopes. They do not establish merge readiness or current-main compatibility.

Historical focus-probe qualification: preserving its old mtime is not fresh-mtime evidence. That specific historical probe is not used as such; independent source hashes and subsequent exact-source frontend checks establish the restored source separately. No source policy or review loop was reopened.

Current-head evidence-only review addendum

Reviewer: /root, independent of the evidence commit author. Exact reviewed PR head: 5c1c225. Delta base/checked production source: 9c1c672. Applied the repository's required review brief, ownership/agreement, organization and evidence/closure rules to this artifact-only delta. This addendum and the source-attributed regional reviews replace the declined external review for their actual scopes; CodeRabbit skipped the Draft and is not reported as an external approval.

Result: no new actionable findings at any priority in this evidence-only delta. Review loop ends. The historical provenance qualification below limits one supporting probe claim; no source fix or speculative review-loop reopening is warranted.

Verified all 25 additions are confined to verification/desktop/evidence/unblocked-audit. No production, tests, schema, dependency, configuration or verification executable changed from the delta base. The four assembled regional identity entries were independently recomputed from Git: zero source differences between each actual checked head and production source9c1. The consumer author's complete Git tree equals production9c1. All desktop and consumer source SHA256 entries, plus the unchanged mode browser inputs, match actual checked-in bytes. Final integration checkout is clean at exact5c1.

Read the evidence index, source identities, author/review reports, JSON outcomes and explicit failed performance table. Regional conclusions retain their actual heads: API review/Rust bd8ef07, native/SDK/OAuth and desktop review4fea9814, final consumer review/frontend9c1. The separate root consumer review includes scoped94/94 and two public ordering probes; overlapping counts are not summed. Full frontend's actual log confirms3998 tests/307files. The two new PNGs were directly visually inspected: retained synthetic draft, no sent message, disabled Tool approval/Ask; sizes21920/31415bytes. Known credential-pattern scan found no raw tokens/private keys, and the reviewed content contains source hashes and harmless loopback URLs rather than ACP capabilities/provider credentials. This scan is supporting evidence, not a comprehensive security certification.

Internal README artifact links resolve in the repository. PR-description relative links were separately found broken in actual GitHub rendering and reported to the coordinator for absolute immutable-head URLs; that body correction requires readback. All five actual GitHub Mermaid diagrams rendered successfully, confirmed by rendered participants/branches and a direct focus-diagram screenshot. The sole root-owned PR-render tab was closed before quiet performance measurement.

Historical qualification: desktop-review.md describes the earlier focus probe restoration as preserving bytes and mtime; that wording does not establish the canonical fresh-mtime condition. Do not relabel this particular historical probe as fresh-mtime proof. The independently recomputed source hashes and later fresh exact-source frontend checks establish the current restored source separately. The current consumer probes explicitly record fresh-mtime restoration. Author selfreviews remain selfreviews; failed mutation runs and historical compilation/abort failures are not counted as clean passes.

Earlier mandatory scripted37/37 and failed performance32/35 remain labeledbd8. The later final-source mode2/2 and mandatory scripted37/37 were verified separately by the browser delegate and belong in a source-attributed PR update, not a relabeling of committed earlier artifacts. Performance budgets are not waived. Quiet final/current-main baseline, whole-browser/retained-app gates, supported-platform CI and merge-result compatibility remain separate outstanding gates. Current upstream88f34e4da is not represented as the older assembled base965797a9f. Seven blocked-by-restructure issues remain excluded; tracker691 remains open. This report is not merge readiness.

No executable tests, compilers, browser measurements or source mutations were run for this artifact-only delta. Lightweight Git/hash/JSON/link checks ended; all root-owned browser/test/compiler processes remain ended. Broader production correctness remains covered only by the explicitly scoped and source-qualified reviews, not by a new whole-code approval in this addendum.

@Varuas37

Copy link
Copy Markdown
Contributor Author

Performance gate disposition: current-source whole performance 33/35 held, failed; only overview-answer fails both layouts, three of three runs each, 50 ms unchanged. Actual current-main baseline fails the same interaction five of five in both layouts. All other final interaction rows including drag pass. This confirms an existing current-main budget failure; it does not prove its source cause, equate it with deferred #669/#710, or waive a merge gate. The grouped fixes remain Draft/blocked while whole functional/retained-app and supported-platform CI continue.

The actual read-only investigation follows; no source edits, browser profiling or speculative performance fix were made:

Read-only overview-answer investigation

Scope: actual production/script path and existing issue relationship, no source edits, compiler, browser profiling or performance rerun. Reviewed integration source9c1 / published5c1 and actual current-main88f34e4dabdeca46486dfbd5294e0a2a00a0a5fa. Git diff confirms the six files named below are byte-identical between main88 and source9c1. This is source-path evidence, not source-level causal attribution.

Measured current-main baseline (delegate): production, headed bundled Chromium, both default layouts, five runs,4xthrottle; actual command/head/clean source in browser-final/current-main-88f34e4d/performance/metadata.json. Calibration passes(4.12 ratio,120ms->120.5ms). Overview-answer fails all five runs/layout: columns max87.9/median76.5ms, sidebar75.6/73.8ms,10 slow frames total. Drag passes all five runs/layout(max42.6/44.4ms). No budget waiver follows from a baseline failure. Final-source full performance is separate and still running when this report was written.

Actual path:

  1. verification/desktop/scripts/perf-budget.mjs:272 opens overview, presses Home, waits three frames, then presses Allow and expects fewer requests after1500ms.
  2. src/desktop/workspace/ui/overview/request-row.tsx:107 resolves the key binding/selected approval option, verifies fresh answer eligibility, then invokes onAnswer.
  3. src/desktop/workspace/ui/overview/overview.tsx:370 answer records the existing settling row and synchronously moves keyboard focus to the next request through focusItem(:224), before dispatching the typed approve/deny command. focusItem calls native focus synchronously; scrolling is already deferred to RAF.
  4. src/desktop/workspace/adapters/store/commands.ts:539 checks the exact visible approval, marks answering, then invokes the injected workspace source.
  5. src/desktop/workspace/adapters/in-memory/in-memory-source.ts:450 approve audits the decision and emits session and transcript changes through the normal listeners. This is the default desktop scenario port used by this UI performance journey, not a live-provider claim.
  6. Overview's admitted answer completion/movement reaches moveOn(:447): flush leaving phase, leaveInPlace(list,item), then flush release. src/desktop/workspace/adapters/dom/overview-reflow.ts:155 copies the actual departing row, reads computed style/offset geometry, appends a positioned ghost; useReflow(:63) reads new item positions in a layout effect when the shape changes and starts transform animations.

Ranked suspects to distinguish with an actual source-level profile: synchronous focus/selected-peek commit; leaveInPlace geometry/style reads following phase flush; layout-effect reflow after row release. The root keydown B_/byte185686 LoAF entry is the delegated event boundary, not sufficient identification of which application read spent the measured36-45ms forced layout. Historical evidence calls the same root event position O_; the minified function name alone does not establish cause. A profile should identify the actual forced-layout call stack and committed layout mutations before a patch. No candidate has been removed or blamed as proven.

Existing issue relationship: #669 is an open performance/follow-up issue labeled deferred, with an explicit exception applying ONLY to PR642. Its historical updated-main-budget.json actually records overview-answer PASS(columnsmax33/median18; sidebarmax50/median35), so these current overview-answer failures are not one of that accepted six-row failure set. Do not claim the old exception applies to PR739. #710 is cursor-owned/deferred pane insertion; #250/#281 are deferred pane drag work. Neither is demonstrated as this answer's source cause. No new issue, workaround label, owner reassignment or deferred implementation was undertaken by this read-only investigation.

Unresolved: source-level causality and legitimate owner disposition for current-main answer failure; actual final-source full performance, full functional/retained-app and CI/merge-result gates. Production source unchanged; root-owned commands ended and no browser/compilers ran during the child's quiet metrics.

@Varuas37

Copy link
Copy Markdown
Contributor Author

CI is a merge blocker on pushed 5c1c2257c630a3f5d5dc83ba422c882ecc598fbd.

Linux and macOS fail native Clippy with warnings denied:

error: use of deprecated method std::sync::atomic::Atomic::<usize>::fetch_update: renamed to try_update for consistency
src-tauri/src/gateway/infrastructure/commands.rs:704:18
-D deprecated implied by -D warnings

This is the added injected settings-fault fixture, not a failed publication assertion. macOS SDK1,099/server2,220 tests passed before that warning. Six earlier jobs pass; Windows remains running. The existing declared MSRV is1.85, so the correction preserves the checked atomic decrement through stable load/compare_exchange rather than introducing a newer API, warning allowance or toolchain pin. Only the test fixture has a working correction; it will not be pushed until the scoped checks and independent review pass. The separate frozen browser production source remains9c1 and its full functional run continues quietly. Draft stays blocked by both CI and the measured frame-budget failures.

Latest CI denies the deprecated fetch_update in the native settings fault
fixture. Keep the declared Rust 1.85 API requirement and the same atomic
checked-decrement semantics using load and compare_exchange. No production
code or warning policy changes.

Regression introduced by 5c949fc (Retain prior settings through
unconfirmed Claude directory publication). Refs #689 and #690.
Deferred overview rows from c13a9ed (#606) are not ready when the
shell appears. The compact-card check introduced in d59d927 (#642)
also reads a frame from before its DOM is visible; rail-fold-focus
from df35d18 (#659) reads resize state before the page commits it.

Use existing publication and frame fences before unchanged assertions.
Individual source removals reproduce the original outcomes; restored
checks pass every default engine, layout and size combination.

Refs #691
7ca52f39c waited for overview publication but ignored the wait's false
outcome. An existing target row could then make semantic assertions
pass while complete-list publication remained unavailable.

Refuse that input through the existing CannotRun outcome. The browser
boundary witness retains real target rows and producer attributes;
removing only the boolean guard reproduces the fail-open result.

Refs #691
Preserve the failed initial whole run, individually failing publication-fence removals and the P2 timeout disposition. Keep final whole/CI and failed performance gates explicit. Refs #691.
@Varuas37

Copy link
Copy Markdown
Contributor Author

Current-head local replacement review for CodeRabbit’s explicit draft decline follows. It verifies the assembled correction bytes and evidence at pushed 05bc3c6c3; completed production reviews keep their original scope and heads. This is a source-qualified review, while CI and the final whole browser run remain active. The 50 ms performance failure still blocks merge.

Current pushed-head replacement review — PR #739

Reviewer /root/unblocked_panel_list, independent of coordinator/integrator. Exact clean pushed/local-remote head 05bc3c6c3b1ab1a9148bfc45f7cb133cc6f855fe; base 5c1c2257c630a3f5d5dc83ba422c882ecc598fbd. Integration checkout /Users/nessa/.codex/worktrees/client-event-delivery/nessa-agent. HEAD and local origin/codex/691-unblocked-audit both equal the brief, and git status is clean. No runtime process, compiler, browser, test or source edit was started.

Result

Zero unresolved findings at blocker, major, minor or nit priority in this explicitly narrow current-head assembly/evidence review. This is a local replacement for the external draft-review decline, not an external CodeRabbit approval or a new whole-code approval at05bc. Earlier regional reviews retain their actual heads. Final assembled browser verification and actual CI remain separate gates; the50ms performance contract remains failed33/35 and is not waived.

Actual source identity and boundaries

Independently checked git objects and actual disk bytes against the committed evidence manifests. Four readiness files at05bc exactly equal the previously reviewed fixed-P2 source 4a3c9a900697d02bfd83e90237c07ae11ed140b0: command-order, drag, columns, canonical UI-load sampling table. All four SHA256 values match both source-manifest and round2-identity JSON. The native commands.rs exactly equals reviewed/tested 31e0ff4ee8c7abed538bb7acaf5bd56d48f63cdb, SHA256 865a30b8890569d80d3127eadb918408e2bf99684bb87bec2f7cdfc8ade7b1da; separately asserted its production prefix before cfg(test) configuration_directory is identical to base.

The full26-file base-to-head inventory is restricted to these five source/doc files plus the necessary unblocked-audit evidence additions. Production frontend, crates, packages, workflow/Cargo metadata have zero diff. The only host file change is the previously reviewed test-fixture expression. No paused issue implementation, additional API policy/state/guard, new producer/marker, fixture, selector, geometry/budget or dependency/path change was introduced by assembly. Existing maps and boundaries remain valid. Exact identity assertions are recorded in /tmp/nessa-unblocked-audit/pr-739-05bc-current-head-review-source.json.

Agreement and evidence attribution

  • Native fixture: current inline checked-decrement CAS bytes retain the original single counter/fault owner, SeqCst ordering, zero behavior, contention retry and inner-write-before-ack-fault sequence, as in the independent source review. Adjacent report/results accurately identify checks against5c1 plus the dirty single-file source hash, local Rust1.98.1 host checks and a load-bearing checked_sub(1)->(0) semantic failure. Restoration records the same hash and fresh mtime. Report explicitly distinguishes older local compiler behavior from actual three-platform deprecated-helper CI failures and requires new-head CI; no old fullRust result is relabelled as execution of this correction.
  • Command readiness: both current callers reject false overviewListed via existing CannotRun before row semantic checks. The prior P2 is closed by current source, the fresh round2 report and its independently labelled evaluation-failure probe: producer/target remain published, but unavailable observation yields two refused cases and zero semantic success. The guard-removal artifact records actual two false passes and witness exit1, not a clean success. No test-only runtime authority enters product code.
  • Drag readiness: current recording generation/reset/retirement remain in unchanged producer/helper bytes; wait consumes only latest owned ghost presence. Original size/pointer/title/residue checks remain intact. Report preserves removal6/8 and restored8/8, with variable failure rows disclosed.
  • Rail readiness: only existing page-frame fence follows viewport resize; current finite-animation settlement,20px descent,400px floor and focus assertions remain. Report preserves removal8/12 and restored12/12; no expected-state wait or budget relaxation.
  • Existing owner fixtures/helpers and product state are not duplicated in evidence. Command metadata accurately targets fixed4a3 source; drag/rail reuse is expressly byte-based from their unchanged clean7ca successful runs. All relative Markdown links in the new evidence overview and scoped reports resolve in the checked tree.

Inspected the current evidence inventory, scoped native/browser author reports, previous native independent report, first readiness P2 report and fresh round2 zero closure; checked author source manifests, exact metadata/source hashes, healthy/negative result attribution and restoration evidence. Committed command8/8, drag8/8, rail12/12 and independent command8/8 positive/0of2 negative are scoped actual results, not whole-run totals. Canonical initial full9c1 failure (20/23 groups held,978/990 rows held), prior perf32/35 atbd8 and later33/35 at9c1 are distinct historical runs. The reports preserve these heads and state the remaining performance failure/whole/CI limits. Historical author/identity snapshots saying review/checks pending are followed by the actual dated-stage closure and execution artifacts rather than rewritten as prior approvals.

Applied current canonical required review brief, agreement across facts/layers, organization and evidence/closure rules. Native persistence/fault relationship and browser DOM-publication/sampled-frame/resize-commit relationships retain their authoritative owners; assembly maps prior evidence to matching bytes without attributing old runtime runs to a new executable head.

Explicit limits and handoff

No completed source review loop was reopened, and no whole-code review at05bc is claimed. This review covers only base5c1->05bc integration identity and evidence consistency, not the original eleven-issue implementation as a new independent pass. No runtime checks, compiler/browser/test process, actual providers, native WKWebView or supported-platform CI were run by this reviewer. The coordinator/browser delegate owns final whole verification at05bc; performance remains a merge blocker under the existing50ms contract even though the current-main baseline also fails. Paused/deferred work is not resumed or used to waive it. No public GitHub mutation was made.

All reviewer-owned work ended after this read-only report; no owned background process exists. Tree remains clean, with no source/generated/fixture edits.

GitHub readback separately confirmed all four body/comment screenshot image elements decoded at 440 × 480 from immutable links. The five existing Mermaid diagrams rendered successfully; their source is unchanged. The only owned PR tab was closed before final quiet browser verification.

@Varuas37

Varuas37 commented Oct 10, 2026 •

Copy link
Copy Markdown
Contributor Author

Current-head CI at 05bc3c6c3b1ab1a9148bfc45f7cb133cc6f855fe has completed: eight owned jobs pass, Windows fails one native concurrency witness, and the required aggregate therefore fails. Windows job.

The native suite reports 427 passed, 1 failed: an_identical_in_flight_directory_change_waits_for_that_attempt, at src-tauri/tests/gateway/application.rs:2467, assertion second.join().unwrap().is_err(). This is separate from the previously corrected deprecated atomic fixture helper; that warning is no longer the failure. Linux and macOS pass. The complete final browser run remains frozen at this same source while the test’s admission ordering is inspected read-only.

The current source increments the existing test-only waiter counter for the first caller as well as the second. The test currently waits only for a nonzero counter before releasing its native gate; that does not establish that the second caller has joined. A later caller can legitimately make a new attempt after the first failure. This is a source-grounded diagnosis under independent checking, not a claimed fixed test or an excuse to rerun the job until green. No source patch, warning suppression, assertion removal or budget change has been pushed for this failure.

PR remains draft. The performance gate is still failed 33/35 against the unchanged 50 ms budget, and final whole verification is still running.

The source permits the following ordering when the test releases on the first waiter. The CI assertion is observed; this exact schedule is the independently confirmed source path, pending a controlled regression witness.

sequenceDiagram
    participant Test
    participant First as First caller
    participant Service as Gateway publication
    participant Native as Held native effect
    participant Second as Second caller
    First->>Service: Claim and wait for owned publication
    Service->>Native: Begin held effect
    Test->>Service: Observe one receipt waiter
    Test->>Native: Release before second caller joins
    Native-->>Service: Refuse publication
    Service-->>First: Settle error and clear prior attempt
    Second->>Service: Enter after prior settlement
    Service->>Native: Begin legitimate later attempt
    Native-->>Service: Later physical attempt succeeds
    Service-->>Second: Settle matching success receipt
    Note over Test,Second: Original assertion incorrectly assumes prior admission
Loading

@Varuas37

Copy link
Copy Markdown
Contributor Author

Final whole browser verification at exact clean pushed 05bc3c6c3b1ab1a9148bfc45f7cb133cc6f855fe is PASS: 23/23 groups, 990/990, exit 0, 3,112.5 seconds (51m52.5s). Chromium and WebKit, all default layouts/sizes, default dev functional mode, no quick or partial functional selection. The wrapper rebuilt its correct private gateway first and includes its separate production message-sync fixture.

Corrected observations hold in the actual complete run: drag 172/172, rail/layout 84/84, command-order 8/8. Focus 80/80, responsive 34/34 (unchanged 400 ms motion contract), safe area 90/90, live mode 2/2, mounted Messages lists 2/2, MCP Apps 92/92 and application review 32/32 also pass. The initial clean9c1 full run remains failed 978/990; these fresh totals are not a relabelling of it.

Production message synchronization passes 14/14, three fresh delivery pages per engine/layout. Chromium is calibrated at 4× throttle; WebKit is unthrottled. Max delivery frame: columns/sidebar 32.6/34.0 ms Chromium, 23/21 ms WebKit, all zero over 50 ms. Sidebar active delivery latency max50.4 ms is against the separate unchanged600 ms delivery contract, not a frame-budget failure. Four dev-only retained-app cases were excluded from the production fixture, not counted as passes; their separate actual dev4/4 result and mandatory scripted37/37 retain source9c1 with exact matching executable-region identity at05bc.

Draft merge blockers remain: separate full interaction performance failed33/35 against50 ms (two overview-answer rows); actual main88 also fails those rows but does not waive the gate. Current-main merge-result CI passes eight jobs including Linux/macOS, while Windows has one test admission-ordering failure and the required aggregate fails. The minimal test-only waiter correction is undergoing its controlled before/fixed witness, then independent review and new-head CI; no repaired Windows result is yet claimed.

All browser-owner processes, dev servers, gateways and temporary caffeinate ended; its private tree remains clean exact05bc. Safe screenshots and compact summary are retained with the source identities. No raw ACP/session capabilities are published, no budget/assertion/fixture is relaxed, and paused restructure issues plus tracker691 remain open.

68f5322 (Keep Claude directory publication owned through native
settlement) made the initiating caller a receipt waiter too. Hold the
existing native/settings test fixture until both callers enter the
retained publication instead of releasing on the first waiter.

Preserve the sole-caller panic/drop gate and original public outcomes.
Refs #689 and #690.
Keep the actual 990/990 whole-browser result at its checked source.
Record the Windows publication-waiter failure, controlled before/fixed
proof, fresh review and complete host checks without relabelling prior
source or waiving the failed 50 ms interaction budget.

Refs #691, #689 and #690.
@Varuas37

Copy link
Copy Markdown
Contributor Author

The Windows test-gate correction is now pushed in the same grouped draft at fb8d621. No production source changed: only three existing native test conditions now wait for both receipt callers, with the existing ADR689 test-fixture row. The fourth sole-caller gate and all original result/settings/host/attempt assertions remain unchanged.

Actual controlled public proof keeps the same scheduler and assertions: before0/1 → fixed1/1 → actual sole-gate removal0/1 → fresh-restored1/1. Before/removal compile successfully, then fail the original second-error assertion with first typed Gateway/Registration failure, second Ok, requested host directory and two attempts. Fixed/restored return the same typed failure to both, restore the host directory to None and run one attempt. Scratch scheduling is absent from the committed tests. Neighboring individual-removal failures are not claimed.

Fresh independent two-file review reports zero findings. Combined full host verification at exact clean source7f9ef1807 passes 705 tests, one existing ignored, 279 tooling tests, six signed-manifest and nine release-key tests, format and all-target Clippy with warnings denied,28.2s. These overlap the author's61-case suite; totals are not added. Proof, source identities, restoration and review.

Final full browser23/23 groups990/990 remains actual source05bc, whose frontend/gateway/verification executable bytes are unchanged by this test-only correction. Its previous failed run stays separate. Fresh current-head CI run38078915630 is running; Windows/aggregate are pending, not inferred from local checks. CodeRabbit skips the draft; a fresh currentfb8 assembly/evidence review is running as the local replacement, with prior production scopes preserved.

Still blocked from merge: production interaction performance33/35, both overview-answer layouts exceed the unchanged50ms contract. Main88 baseline failure neither proves the source cause nor waives the gate. Seven restructure-labelled issues remain paused; all eleven active issues remain In review and tracker691 remains open. No merge/ready/closure is claimed.

@Varuas37

Copy link
Copy Markdown
Contributor Author

Fresh current-head local replacement review for the explicit draft-review decline, exact pushed fb8d62152d2c26c9ae6e659b7b02d7f5cadfff38, follows. This source/assembly review is independent of the native gate author and prior correction reviewer; production regions and runtime checks retain their actual heads. Performance still blocks merge, and fresh supported-platform CI is pending.

Current-head assembly and evidence review — PR739

Reviewed clean fb8d62152d2c26c9ae6e659b7b02d7f5cadfff38 in /Users/nessa/.codex/worktrees/client-event-delivery/nessa-agent, against 05bc3c6c3b1ab1a9148bfc45f7cb133cc6f855fe. Fresh reviewer: browser-verification owner, different from the native-gate author and prior coordinator reviewer. This is the local current-head replacement for the draft external-review decline; a declined external review is not a pass.

Result: zero new unresolved findings at any priority in this review's assigned source/assembly/evidence scope. No blocker, major, minor or nit finding was dismissed. The separate known performance gate remains FAILED33/35 against50ms and still blocks merge; fresh supported-platform/current-head CI remains required. This review does not approve merging the draft.

Inventory and source identity

Independently read the complete base-to-head diff:29 files, two source/document changes and27 evidence/index files. The only executable-source changes are the three existing waiter conditions in src-tauri/tests/gateway/application.rs at2457,2529,5065. The existing ADR689 ordering table gains one test-fixture-release row at46. No production, frontend, gateway implementation, client, SDK, protocol, configuration/workflow, verification script/fixture/helper or tooling executable changes occur. No paths or module boundaries move, so no module-map adjustment is needed. The evidence folders extend the existing unblocked-audit index rather than adding another contract authority.

Actual SHA256:

  • application test: f2afceaf863adfc94731932e30deaba1ca26af934465eb26531d0c05b5eebd61.
  • ADR689: 27802751ccc2f45fed0b134a542f757e622e0b1e676970571a9c6ea9c1db6a4e.

Both match the committed final-source manifest and are byte-identical at authorb323 and reviewed integration7f9. Currentfb8 adds evidence only beyond7f9. The sole-first-caller gate in application/claude_publication.rs:352 is unchanged. The final test file contains exactly three <2 gates and no temporary scheduler/debug scaffold. Final identity/inventory is in adjacent pr-739-fb8-current-head-review-identity.json.

Related facts and their enforcers

Traced the changed tests through actual Gateway::change_claude_configuration, claim_claude_directory, wait_claude_publication, ClaudePublication::wait, ClaudePublicationFinish::complete/restore_failed/Drop and the held host/settings fixtures. The publication slot and retained Arc identify the transaction. Its initiating caller now consumes that same receipt; the existing cfg(test) per-publication counter increments before its result wait. Same-directory callers join the pending publication, while different-directory callers consume its settlement before claiming their own transaction. Each changed test creates exactly two callers. With the native/settings effect held, the outcome cannot settle or replace that publication. A count of one therefore does not establish second receipt admission; <2 establishes both consumers before fixture release. It does not authorize production effects or introduce another owner.

Checked these facts together with returned outcomes and authoritative state: identical callers share one held registration failure and restore the directory with one native attempt; different callers wait before the permitted later success with two attempts and matching durable/live directory; the unchanged durable save holds the next change until settings acknowledgement, then both accepted saves complete. Original assertions, deadlines and pacing are unchanged. The sole-first-caller panic/drop gate correctly still asks for one consumer. Adjacent dropped/panicking caller and pending-result paths remain source-unchanged, rather than being inferred from a counter alone. ADR689 already names689/690 ownership and records the new fixture ordering in its existing table; no paused/history ownership or new production policy is introduced.

Evidence and claims

Read both new evidence folders, relevant execution logs, controlled scheduling patch, mutation/restoration manifests, source identities and independent report. The injected scheduling witness is explicitly test-local: second public API entry is held until the existing waiting-loop body admits it, or after first settlement if the old gate skips that loop. Actual typed results and directory/attempt state distinguish the original assertion failure from compilation failure. Before/removal exit101 and fixed/restored exit0 match the recorded one-selected-test logs. Fresh-mtime restoration and source occurrence/hash records are present. Auxiliary zero-selected binaries and overlapping service/full-host suites are not counted as extra passes. Neighboring individual removal failures are explicitly unclaimed, rather than invented.

The independent complete host check is correctly attributed to7f9 (same source bytes asfb8):279 tooling tests,705 host tests plus1 existing ignored,6 manifest and9 release-key tests, formatting/all-target Clippy, exit0 in28.2s. Local macOS/Rust1.98.1 support is not relabeled as Windows/MSRV execution. The historical actual05bc Windows427/428 failure remains failed; the correction explains the reachable admission ordering without claiming its injected schedule was the recorded Windows instruction trace. Renewed supported-platform CI is explicitly still required.

The whole browser summary is byte-for-byte the actual owned final05bc summary:990/990,23/23 default groups, exit0,3112.5s. Relevant production/browser executables are unchanged atfb8, but this evidence remains an actual05bc run, not a freshly executedfb8 whole approval. Separate scripted37/37 and retained-app4/4 remain actual9c1 runs with declared byte-identity reuse; four dev-only rows excluded by the internal production fixture are disclosed. Initial9c1 whole978/990 failures and readiness removal/P2-closure evidence remain preserved. The old05bc merge/CI identity and image/diagram readbacks retain their original heads and running/pending meanings.

Performance retains production9c1 source,33/35 failed rows, original50ms thresholds and LoAF attribution. The selected baseline is explicitly actual88f34e4d, not later2d2a main. No deferred669/710 causal fix or budget waiver is claimed. Source delta imports none of paused678/680/681/682/683/686/696; active source-review boundaries remain unchanged. Existing historical evidence/index entries are clearly qualified by initial publication or their actual source head, with latest checkpoints linked separately.

Checked34 local Markdown file links in the changed inventory: zero missing targets. All13 new JSON files parse;23 group totals sum to990. No raw relay capabilities, ACP payloads or credential traces appear in this new evidence inventory. The evidence scaffold is a labelled patch artifact, absent from compiled source.

Limits and process

Applied canonical current required-brief, local review, agreement, evidence/closure and organization rules. This review is read-only source/evidence inspection: no compiler, test, browser or GitHub operation was started, no source was edited, and the current checkout stayed clean exactfb8. I did not repeat the native injected scheduler/full-host suite, reopen completed API/UI review loops or assert repository-wide absence outside the scoped region identities. My earlier full real-browser execution is preserved at its actual05bc head. Native WKWebView, live providers, Linux installation, packaged signing and new supported-platform CI remain unclaimed. The performance failure is an unresolved merge gate, independently of this zero-finding correction review.

@Varuas37

Copy link
Copy Markdown
Contributor Author

Current pushed head fb8d621 now has all10 owned CI checks passing, including Linux, macOS, Windows and the required aggregate. Run38078915630. The actual checked merge is db98edc8718f52b8328880358feea5d1cd8d3678, parents current main2d2a and PRfb8, independently verified from the backend workflow checkout.

Backend Windows job explicitly reports:

an_identical_in_flight_directory_change_waits_for_that_attempt ... ok
test result: ok. 428 passed; 0 failed; 0 ignored; 0 filtered out

Six manifest and nine release-key tests also pass. Windows log output is truncated; the actual job conclusion and untruncated required aggregate both independently report success. Existing supported-platform Unix-only omissions remain their recorded limitations, not new passes. The two earlier CI failures are preserved at their original5c1/05bc sources rather than relabelled green. CodeRabbit still skips the draft; fresh current-head source-qualified local replacement review is posted, not an automated external approval.

Final functional verification remains actual 23/23 groups990/990, plus identity-qualified scripted37/37 and separate retained dev4/4.

This PR remains OPEN DRAFT, blocked from merge solely by the still-failed interaction performance gate33/35: overview-answer columns76.8/73.3ms and sidebar72.7/65.5ms max/median against50ms, all three trials in each layout slow. Actual measured main88 also fails those rows but neither supplies a source cause nor waives50ms. No deferred669/710 scope or historical642 exception is used. The next step is a source-level profile of the forced-layout call stack before an owned correction; no source cause or performance fix is claimed here.

All eleven unblocked fixes stay In review; seven restructure-labelled issues stay paused/excluded in Backlog; tracker691 remains OPEN. No ready/merge/closure occurred. All owned local processes/tabs are ended and the integration tree is clean.

@Varuas37
Varuas37 marked this pull request as ready for review October 11, 2026 06:10

@greptile-apps greptile-apps 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.

Your trial has ended. Reactivate Greptile to resume code reviews.

@chatgpt-codex-connector

chatgpt-codex-connector Bot commented Oct 11, 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 2026-10-11T06:16:09.996309Z fb8d621 Draft marked ready
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.

@Varuas37
Varuas37 merged commit 8bafbfc into main Oct 11, 2026
10 of 11 checks passed
@Varuas37
Varuas37 deleted the codex/691-unblocked-audit branch October 11, 2026 06:11
Varuas37 pushed a commit that referenced this pull request Oct 11, 2026
One conflict, in the conversation projection tests: this branch's
artifact tests and main's selection and runtime revision tests are
independent, so both are kept. Every file both sides touched carries
this branch's change unaltered, and the generated protocol files match.

Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01HFdiN79wR9Q5CcAFx7UYP7
Varuas37 added a commit to nessalabs/skills that referenced this pull request Oct 17, 2026
## Motivation

The recovered October9–12 ADR audits are the remaining source evidence
deferred by [#23](#23). Most of
their lessons already live in the ADR reconciliation procedure added by
[#22](#22): decision acceptance
is separate from delivery, slices need their own evidence, inventories
need lifecycle placement, and stale claims must be repaired wherever
readers encounter them.

One useful distinction can be clearer: a feature name is not proof that
the shipped design satisfies the decided contract. A different mechanism
may preserve the guarantee, leave a real implementation gap, or require
an explicit decision revision. This draft makes that comparison explicit
without prescribing the historical proposal as the correct design.

## Solution

1. `8a72865`: amend the existing ADR scope-reconciliation paragraph to
compare promised guarantees and distinguish an implementation gap from a
decision needing explicit revision.
2. `8a72865`: add one Core Rule2 example that links to that owner; keep
all historical project detail here.

Only two existing documentation files change. This is a narrow
clarification, so package/plugin versions remain `0.6.0`; the repository
requires bumps for broad changes. No new skill, duplicate audit
checklist, workflow, implementation, ADR file or issue is introduced.
**Leave this PR as a draft; do not merge without separate approval.**

### Baseline and source access

- Based on the actual default `wip` at
`572b154685a75892b54d939e1d2d9b0e6fb22477`, the merged #23 tree. Root
`AGENTS.md`, the engineering ownership table, ADR, architecture, method
and review owners were inspected. No additional applicable `.agents`
instructions were present.
- No competing open skills PR was found before editing. Earlier skills
issues are older Rust research, not a new ADR integration.
- Input is the recovered historical report-message evidence supplied
with this task. Full thread text was not accessible in this executor;
ZIP issue-body attachments remain unavailable. Neither original issue
drafts nor publication receipts are claimed inspected.
- Read-only public source checks inspected relevant PR metadata and
excerpts of the ADR index, runtime ADR, Read-grants design and
physical-persistence design at the October12 snapshot. These corroborate
the bounded mechanism comparison; they are not a fresh audit of current
Nessa.

### Source-to-owner dispositions

| Recovered finding family | Disposition | Existing owner / reason |
| --- | --- | --- |
| 501 linked once with unique identity but absent from lifecycle table |
Already covered | ADR inventory structure requires repository-consistent
placement and says a stray link is not index coverage. Uniqueness and
membership are separate checks already expressed. |
| 252 acceptance versus implementation checklist and tracking state |
Already covered | ADR scope reconciliation separates accepted decisions,
shipped slices, evidence and remaining gates; reader-surface repair
includes tracking issues. |
| Partial replay, grants, artifacts, pairing and outbound delivery |
Already covered | Per-slice evidence and gates; one merged
implementation does not establish full scope. |
| A proposed Cedar/semantic-record design versus the shipped metadata
grant table and durable journal | **Small amendment** | ADR scope
reconciliation now requires matching the promised guarantee instead of
the feature name and distinguishing implementation gaps from explicit
decision revisions. It does not decide which design is right. |
| Stale 0008/0009/202/302/392 prose, index/body conflicts and 688/689
status | Already covered | ADR contradiction repair covers header,
current prose, tables, checklists, index, README and tracking issue;
placement follows the repository convention. |
| Persistence mechanics and 27 ordering cases without decision
rationale/rejected alternatives | Already covered | ADR template
requires the binding constraint, alternatives and reasons; method's
decision artifact also includes rejected options. Implementation
instructions and reported tests are not newly relabeled as decision
history. |
| PR715 already addressed replay documentation; avoid filing it again |
Already covered | ADR's existing issue/open-PR comparison checks the
exact changed sections, not document overlap. No Nessa issue action is
taken. |
| Repeat-audit baseline/delta and outcome separation | Coverage
practice; no additional skill rule | Pin historical revisions and report
new/still-open/reconciled findings in this ledger under existing
evidence-class rules. A publication failure is not proof inspection
never completed. |
| Reported publication failures and disabled ADR automation | Reported
operational outcome only | No independent automation-state verification,
schedule change or publication retry. Rust research state is outside
this patch. |

### Recovered-report coverage ledger

These are historical findings at the stated revisions, not claims that
the defects remain in current Nessa. Repeated themes are counted once as
transferable guidance.

| Report message | Historical Nessa snapshot | Recovered coverage and
disposition |
| --- | --- | --- |
| October9 · `2a086b87-e13b-43af-a099-4a2f577bcd82` |
`4d1b7278cfe633a6c38094f0bcfb45d4d5043af0` | Reported 36 records:23
done/13 todo;326/344 moved done,252 accepted with implementation
separately tracked. 0008,202/302,392,501 and persistence-rationale
findings map to existing owners; no duplicate replay-doc work. |
| October10 · `2d3f1bcd-f54a-4d96-b836-3caa1a19e0b9` |
`88f34e4dabdeca46486dfbd5294e0a2a00a0a5fa` |501 membership,715 stale
replay index and720 unchecked grants slice are already covered. The
proposed-versus-shipped semantic-contract distinction supports the sole
amendment; peer738 and persistence rationale remain historical case
evidence. |
| October11 · `e4ad69af-9798-4d6f-a666-3a3dbb8aad70` |
`bae7a073b597285823b92f037d7079cf4522a716` | Reported 38 records:23
done/13 todo/2 root;688/689 still under review after739,744 pairing
shipped while748 outbound was not yet merged at this snapshot. Existing
lifecycle/slice/revision rules cover these distinctions. |
| October12 · `a452f573-d4e0-4663-a720-3970e6313f15` |
`07799e6405812a52297c02bac39525ec0315568e` | Reported37 further commits
and38 records; grant/artifact/lease work advanced while checklists
lagged;392,501 and rationale gaps persisted. Already-covered themes
reinforce the narrow clarification. Reported ADR automation disabling is
separate from audit completion and Rust research. |

Provenance: recovered messages belong to conversation
`6aba0796-656c-83e8-a735-49b9ed36b026`. The identifiers preserve source
attribution without putting private conversation links into reusable
skills.

**Unavailable attachments:** issue-body ZIP attachments were not read,
so exact unpublished drafts, their complete acceptance language and
publication outcomes were not independently reconstructed.
**October13–16:** no report messages were found by the recovery; these
dates are not counted as inaccessible completed reports. There is no
basis here to invent additional audit runs or claim a current Nessa
audit.

### Public supporting evidence

- At the [October12 runtime ADR,
sliceG](https://github.com/nessalabs/nessa-agent/blob/07799e6405812a52297c02bac39525ec0315568e/docs/adr/todo/252-runtime-roles-and-execution-leases.md#implementation-plan),
the wording names a Cedar grant. The [same-revision Read-grants
design](https://github.com/nessalabs/nessa-agent/blob/07799e6405812a52297c02bac39525ec0315568e/docs/design/read-grants.md#what-this-slice-does-not-do)
explicitly distinguishes its metadata table/journal from in-stream
semantic records and policy-bundle changes.
[#720](nessalabs/nessa-agent#720) is the
delivered slice. This is evidence to compare contracts, not proof that
every mechanism difference is a defect or that a chosen implementation
must be reverted.
- The [pinned ADR
index](https://github.com/nessalabs/nessa-agent/blob/07799e6405812a52297c02bac39525ec0315568e/docs/adr/README.md)
and [#715](nessalabs/nessa-agent#715),
[#739](nessalabs/nessa-agent#739) support the
historical projection/lifecycle examples; those rules already exist.
- [#744](nessalabs/nessa-agent#744) merged at
the October11 report's exact snapshot,15:43UTC.
[#748](nessalabs/nessa-agent#748) merged later
at16:37UTC. “Not yet merged” is a snapshot observation, not a current
outbound-delivery defect.
- [#652](nessalabs/nessa-agent#652) and its
[pinned physical-persistence
design](https://github.com/nessalabs/nessa-agent/blob/07799e6405812a52297c02bac39525ec0315568e/docs/design/bounded-physical-persistence.md)
support the distinction between implementation mechanics/reported tests
and a decision's rationale. This draft makes no new current missing-ADR
claim.

### Validation and limits

- **Run:** repository-wide local Markdown path and heading-anchor
validation:182 links pass (fenced template examples excluded).
- **Run:** all six declared skill paths, required frontmatter and agent
descriptors checked; package/plugin versions both remain `0.6.0`.
- **Run:** `git diff --check`; refreshed `origin/wip` remains at the
stated base before publication.
- **Run:** author diff review and independent source-to-owner triage,
followed by independent review of the final diff and this description.
No material blockers, duplicate owner or evidence/coverage overclaims
found.
- **Read-only source checks:** the public PRs and pinned documents
linked above resolved; selected historical excerpts were inspected. This
does not establish runtime behavior or an exhaustive current audit.
- **Unavailable:** original full thread text in this executor and ZIP
issue-body attachments. The supplied recovered-message evidence is
distinguished from source excerpts inspected here.
- **Not run:** runtime/native tests, historical ordering cases,
benchmarks, installed-skill behavioral evaluation or production defect
reproduction. No CI workflows or standard checked-in skill validator
exist in this repository; the documentation checks above were run
locally.

No `nessa-agent` files, issues, PRs or schedules were changed; the
separate #953 task was untouched. Prior denied issue publications were
not retried. No native/runtime tests, ordering cases or historical audit
checks were rerun. The change teaches a review distinction; it does not
assert a runtime-validated result or repair historical Nessa defects.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment