[david6666666]

Refs #314

What changed

• Keep summary links to recorded runs enabled when their outcome is not yet recorded; retain the warning and keep missing, corrupt and ambiguous run links disabled.

• Distinguish a recorded candidate from missing or corrupt run evidence, and warn when a summary cites a run before its outcome is recorded.

• Bound synthesized branch-chart scales and shared axis ticks; withhold unsafe charts while retaining recorded data. Preserve candidate IDs already present in the campaign and single numeric GPU ordinals when rendering redacted commands. The GPU allowlist intentionally accepts one 1–3 digit ordinal, such as --gpu 4; comma-separated selectors such as --gpu 0,1 remain redacted.

• Remote viewers connect through an SSH local port forward to the server on the remote machine. The server always binds to loopback and has no non-loopback binding option; it does not add authentication for other users on that machine.

• Compress live HTTP responses and intern repeated trace contexts during transport; the browser restores complete text and existing uncompressed API consumers remain supported.

• Keep the live server responsive when a browser opens an idle speculative connection. Mark cancelled runs separately from execution failures, retain their evidence on the timeline, and translate the state in Chinese.

• Restore the performance line for comparable independent candidate measurements with unique round numbers. The chart clearly identifies non-cumulative results, preserves quality failures and rollbacks, and retains explicitly recorded cumulative charts.

• Add optional live observation with inferrsi dashboard --input <RUN_OR_CAMPAIGN> --live --port 8765: a read-only loopback server supplies recorded updates about every three seconds without a page reload. The page retains selections where possible, shows refresh failures and retains its last view. Offline export remains available.

• Address Dong1017's translation review: caller overlays apply only to explicitly marked narrative, while verdicts, quality and missing-evidence states use code-owned translations. Add regression coverage for hostile overlays in both languages.

• Bind story logs and knowledge captures to their campaign and capture IDs; check references against the captures actually named by each round. Withhold corrupt or mismatched records and keep authored narratives separate from technical stage events.

• Add an offline English/Chinese language switch, with a reviewed caller-supplied translation overlay for recorded narratives and access to the original knowledge context.

• Export a self-contained, offline optimization dashboard for existing runs and explicitly recorded campaigns. Candidate details bring together the change, code or configuration locations, benchmark and accuracy methods, redacted commands, and measured evidence.

• Show a caller-reviewed optimization trace with the assembled knowledge context, search queries and sources, declared search space, ordered observations, attempts, decisions, rationale, and success and failure summaries. The separate Selected optimization and Stage coverage sections have been removed.

• Add a bounded trace snapshot format and a CLI operation that inserts an explicitly reviewed context file without silently truncating it. Candidate, optional run and round identities, source file hashes, and explicit event references are checked; missing, redacted, corrupt, mismatched and ambiguous records remain visible as such.

• Add a campaign-level Optimization story after the two performance charts, so a reviewer can follow the whole optimization rather than one candidate at a time: the goal; each knowledge retrieval and why it was made; the declared search space with options marked tried, pruned or untried and the reason; every round's hypothesis, why this optimization was chosen, decision and next step; and success and failure summaries that cite evidence and name the failing stage. Views: Optimization process, Knowledge context, Search space, Successes and failures.

• Add inferrsi campaign capture-context. It runs the same collect operation the agent uses, with a context budget, and keeps the complete envelope: the request, what search selected, what was left out of the context and why, and every resource body with its SHA-256. The assembled text is re-derived when the dashboard loads, so an edited, missing or mismatched capture is withheld rather than shown. This replaces hand-supplied context text for campaigns; the per-candidate snapshot above is unchanged.

• Add inferrsi campaign log, which validates and stores a numbered, versioned campaign optimization log (schema_version: 2). Earlier versions are kept; the latest is shown only when its hash and campaign identity match. Validation errors name the offending field.

• Cross-check the narrative against evidence. The dashboard resolves each reference in the log against captures, rounds.jsonl and recorded runs and lists disagreements under Narrative vs evidence: knowledge cited but never retrieved, or selected by search but dropped by the context budget before it was included in the assembled context; a decision that differs from rounds.jsonl; an adopted candidate whose recorded outcome is not an improvement; a search option marked tried that no round tried; a success whose evidence did not improve; and a recorded failure, regression or rollback that no failure summary cites.

• Keep technical runner and agent events searchable, and allow a narrative step to open a linked event only when its source, line and source-file hash match. A candidate without a recorded run cannot read evidence from a placeholder directory.

• Retain the charts, diagnostic evidence, local review export, and conservative command filtering. The host agent owns context capture and search decisions; the dashboard does not intercept unrelated agents or independently verify caller narrative.

The requester-directed scope amendment records the later bilingual, optional loopback live, and caller-authored story requirements alongside the original offline scope. Raw logs, environments and unreviewed agent requests remain outside the export.

Validation

• Summary navigation: 109 targeted tests and Ruff passed on Windows and remote Linux. Synthetic browser checks verified selection and the missing-outcome warning in English and Chinese.

• Missing-run review fix: 158 targeted tests passed on Windows and remote Linux; Ruff passed. Synthetic browser checks covered missing, corrupt and running-without-outcome cases in English and Chinese.

• Chart-scale and command-display changes: 123 targeted dashboard, campaign and summary-chart tests passed on Windows and remote Linux; Ruff passed. The live API was tested with extreme positive and negative values.

• Latest live-server, transport and cancellation changes: 66 targeted dashboard tests passed on Windows and remote Linux; ruff check src tests tools passed. The Chinese real-time browser view was checked against an in-progress campaign; incomplete attempts remain inconclusive.

• Round trend: targeted Windows tests and Ruff passed; 125 targeted tests and Ruff passed on remote Linux. The Chinese ten-round live recording uses explicitly synthetic data.

• Live mode: tested incremental record updates, conditional responses, local-only binding, invalid Host rejection, source-file isolation, refresh failure and recovery, and offline CSP. Browser checks covered automatic updates, filter and language preservation, and failure/retry. The PR comment contains an accelerated, synthetic browser recording.

• Windows: 153 targeted dashboard, campaign, trace, plugin and CLI tests passed; ruff check src tests tools passed.

• Remote Linux at 806e71d: 123 targeted live dashboard, campaign, trace, story and language tests passed, and ruff check src tests tools passed. One known pre-existing nested-JSON error-message assertion was excluded; it also fails on the unchanged baseline.

• Optimization story: 198 targeted tests passed (dashboard, campaign, trace, plugin, CLI, collect and knowledge CLI, including 18 new tests in tests/test_optimization_story.py); ruff check src tests tools passed. Full suite on Windows: the only errors are in test_mlc_llm_repo_to_skill.py, test_mlx_swift_lm_repo_to_skill.py and test_anthropic_edit_agent.py, caused by Windows limits (environment variable longer than 32767 characters, GBK default encoding, no os.mkfifo). They also error on the unmodified PR head, though not with identical counts; they need a Linux CI run.

• Earlier browser checks covered all four story views, candidate links, context originals, both languages, and the position of both charts. A hostile translation overlay could not change FAIL, INCONCLUSIVE, quality, or missing-evidence labels. The screenshots use synthetic campaign data; the demo's knowledge capture is packaged-library collect output.

• Earlier independent review covered trace privacy and identity binding, path replacement, missing runs, event-hash navigation, old records and rendering. Browser checks covered real and explicitly synthetic snapshots, language switching, candidate selection, context originals, localized review export and responsive layout.

• The story's narrative is caller-authored. Capture hashes check internal consistency, not authenticated origin or model consumption; the dashboard checks references and recorded outcomes, not the soundness of the reasoning. Chinese support covers interface text; log prose appears as written unless a translation overlay is supplied.

• The current head received GitHub technical approvals, including the Amy review. Direct requester confirmation or a link to the original instruction for the expanded [Task] RSI Dashboard: Optimization Workflow Visualization and Trace Capture #314 scope remains outstanding.

• Existing records cannot reconstruct a context that was never saved. Captured text must be reviewed before recording and sharing. Existing candidate records do not yet have a CLI operation to replace their trace reference. This is a snapshot, and a recorded validation command does not prove untested quality properties.

Consult the current-head GitHub checks and reviews below. Technical approval does not decide #314 scope or campaign acceptance.

Screenshots

All screenshots use clearly labelled synthetic data.

🤖 Generated with Claude Code

[chatgpt-codex-connector]

Codex Review Summary

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

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-11-02T12:12:55.676911Z	b4e06aa	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.

[Dong1017]

LGTM for the offline dashboard and correlation implementation at b4e06aae.

• I independently traced the CLI, campaign/candidate/decision records, opt-in agent envelope, evidence reader, charts, local Markdown export, and wheel/plugin packaging. I found no blocking defect in this bounded software change.
• Missing, corrupt and unlinked evidence stays explicit; incompatible or unsupported cross-run comparisons are withheld. Failed quality checks and recorded rollbacks cannot become the best gain. The documented caller-serialized writes and snapshot/privacy limits match the implementation.
• Current-head Ready checks pass, including Python 3.10/3.11 full checks and installed packages, PR references and KV qualification. Knowledge admission is Not applicable. Earlier cancelled Draft checks are superseded by the successful Ready run.
• Existing owner @david6666666 retains integration and the remaining Issue #314 acceptance work. This software approval does not establish campaign effects or independently verify the private real-run/browser evidence.

Evidence and review scope

Reviewed head: b4e06aaef10ea6ba073b1312bd92b5f3eb1a748c.
API base and observed main: 2baebf6ebb815ebf07d21c41e3b3754858c0798e.
Scope: the 17-path diff, affected RunSpec/runner and agent-protocol consumers, regression-test definitions, and the recorded owner scope. This maintenance session did not implement the PR.

Verification: static source review and existing Ready run 41442538882, plus current per-head status and independent workflow inspection. I did not execute tests, a browser, models, devices or provider calls. Hosted checks are software evidence; author-reported demonstrations retain their own limits. No merge or Issue completion action is included.

[Dong1017]

LGTM for the bounded software delta at 0c1cec64, continuing the previous review of b4e06aae.

• Inspected the eight-file delta and affected runner/CLI consumers and regression-test definitions. Caller-recorded candidate descriptions/commands remain evidence inputs, not measured benefits; structured traces stay allowlisted, corrupt/unlinked records remain explicit, and the views render text rather than injected HTML.
• Current-head Ready checks passed on both Python versions, including clean-wheel checks, references and synthetic KV qualification. No new actionable blocker found in this delta; the previous review's scope/limits remain applicable.
• David retains integration and Issue #314 acceptance. This review did not execute code, tests or a browser, and does not independently verify private real-run/browser evidence or establish campaign efficacy.

Revision evidence

Head: 0c1cec6. Base: 2baebf6. [Inspected delta](https://github.com/InferMatrixLab/InferRSI/compare/b4e06aaef10ea6ba073b1312bd92b5f3eb1a748c...0c1cec6). [Recorded scope](https://github.com//issues/314#issuecomment-6575827878). COMMENT only; no formal approval, merge or Issue-completion action.

[Dong1017]

Incremental static review of the new bilingual dashboard at 6b611821: one P2 evidence-display integrity risk.

The new caller-supplied translation overlay is currently applied to the dashboard's fixed verdict/status presentation as well as recorded narrative. The inline note gives a concrete collision case and the smallest safe separation. The original evidence data stays unchanged, but a misleading visual verdict is still possible. Please keep caller-reviewed narrative translations separate from evidence-derived states.

The English-only scope versus bilingual addition is already explicitly tracked in Issue #314's current status; its existing owner retains that scope decision and acceptance. I reviewed the new localization path, relevant Python translation loader and regression tests, and prior reviews; I did not execute tests or browser interactions. No merge or acceptance decision is implied.

[Dong1017]

P2 | A translation overlay can relabel evidence verdicts as their opposites. translated() prioritizes (data.translations || {})[language] for every string. localizePage() walks all ordinary DOM text (lines 321–328), including the pill(row.verdict) / quality / state cells created by renderRows() (lines 399–400) and the verdict in renderCandidate() (line 421). With an otherwise schema-valid reviewed overlay such as {"schema_version":1,"zh":{"FAIL":"PASS","failed":"passed"}}, switching to 中文 visually presents a failed candidate as passed, although the underlying JSON still says FAIL/failed. This is a conditional display-integrity risk, not a claim that the stored evidence is mutated. Please restrict supplied translations to explicitly identified narrative fields (or shield evidence-derived badges/labels and use only fixed, code-owned localized status mappings), and add a collision regression covering FAIL / INCONCLUSIVE, quality and missing-evidence labels. The translation map should never be able to change the meaning of a canonical verdict.

[Dong1017]

Incremental static review of the new optimization-story provenance checks at d7a413e8: one P2 round-scope consistency finding. The inline comment gives a concrete two-capture scenario where globally available knowledge masks an omission in the capture explicitly associated with a round, plus a minimal regression/fix direction.

Scope: inspected optimization_story.py, the capture envelope/verifier, optimization-log schema, related tests and dashboard contract. This is a code-reading conclusion; I did not execute tests, a browser, model or hardware workload. The previously raised translation/verdict collision remains in its original thread and is not repeated here. This COMMENTED review makes no acceptance, CI, merge or owner decision.

[Dong1017]

P2 | Round-scoped knowledge provenance can be mistaken for a globally available capture. captured_refs combines each capture's materialized refs with search-selected refs (line 81), while read_refs is global across all captures (line 90). For example, capture A contains the body for package:k, but capture B only selected package:k and omitted its body due to the context budget. An iteration with context_ids=[B] and knowledge_refs=["package:k"] passes unread() because allowed ∩ read_refs contains package:k via A. The dashboard therefore misses the expected “selected but omitted from this round's context” warning. If a named capture yields no refs, allowed or None also falls back to global scope and can suppress the warning entirely.

Please retain per-capture materialized and selected/omitted ref sets, then check iteration citations against the captures explicitly named by that round. Preserve a distinct empty named scope rather than silently converting it to the global scope. A two-capture regression (A contains a ref, B selects but omits the same ref; the round names only B), plus an empty/invalid named-capture case, should exercise this. This is a static provenance-check issue, not a claim that the source capture hashes or underlying measurements changed.

[Dong1017]

The two previously reported P2 paths have corresponding fixes at 870e752f; this is a scoped repair confirmation, not full-PR approval.

• Verdict display: fixed UI/evidence text now uses code-owned labels; caller overlays are restricted to explicitly marked narrative fields. The regression definition covers hostile FAIL/INCONCLUSIVE, quality, missing-evidence and decision mappings in both languages.
• Round provenance: materialized and selected refs are retained per capture. A round naming a capture no longer borrows another capture's materialized text or falls back globally when that named capture is unavailable. The two-capture and unavailable-capture regressions cover the original scenarios.
• Current-head Ready and declaration checks pass. This feedback comes from source and test-definition inspection plus existing hosted results; I did not execute tests, a browser, model or hardware workload.
• david6666666 retains the remaining story/trace review and scope reconciliation in Issue #314. The added bilingual/story scope and whole-task acceptance remain separate; I have not resolved either historical review thread.

Evidence and scope

• Reviewed head: 870e752f0ea82263bf521a402b7c0e408e829f48; API base: 2baebf6ebb815ebf07d21c41e3b3754858c0798e; assessed main: 663f3b327998526a13aa31e72ad729025c18983c.
• Original findings: translation/verdict collision, round-scoped capture provenance.
• Narrative-only overlay and fixed-label routing; per-capture materialized/selected refs. Also inspected capture-envelope verification and relevant regression definitions.
• Current Ready run, declaration policy. Hosted success is software evidence, not target acceptance or proof of private/browser/device effects.
• Scope is the two repair paths above. The full new 25-file story/trace/localization scope has not received a new independent approval from this maintenance review.

[david6666666]

Live dashboard demonstration

The accelerated browser recording below uses an explicitly synthetic campaign. As its demo writer records a running candidate, a completed result, a failed attempt, another candidate and an optimization log, the dashboard updates without a page reload. The selected candidate remains selected while new records appear.

All figures and candidate results in the video are fictional; this is a UI demonstration and makes no performance or quality claim.

To observe a recorded run or campaign locally:

inferrsi dashboard --input <RUN_OR_CAMPAIGN_DIRECTORY> --live --port 8765

The read-only server binds to loopback, checks for recorded updates approximately every three seconds, and retains the last view if a refresh fails. Existing offline HTML export remains available. The live page displays only data the experiment has actually recorded; it does not capture an agent's reasoning automatically.

inferrsi-live-dashboard-demo.mp4

[Dong1017]

Scoped static review of the live-dashboard delta at 806e71dd: no new actionable defect found in the paths inspected.

I inspected the eight-file delta, including CLI flag handling, the loopback server/Host and path restrictions, snapshot and chart generation, poll/selection/render handling, the regression definitions and the documented local-access and non-transactional-read limits. Ready CI passed on this head.

This is a source-and-existing-CI review of the live delta, not a full renewed review of the story/trace additions, a formal approval, or acceptance of the new live scope. I did not run tests, a browser or an experiment. The server is unauthenticated to local users as documented; snapshots across concurrently written files are not transactional. Issue #314's owner retains scope reconciliation and acceptance, and the earlier two repair confirmations remain limited to their cited findings.

[david6666666]

Ten-round live trend demonstration / 十轮实时趋势演示

The Chinese-language video is approximately 2× speed. It starts with an empty campaign and shows the dashboard refreshing as ten simulated candidate records are written. All candidate changes, measurements, quality results, and percentages in this recording are synthetic; they are not GPU benchmark results.

Each point is an independent candidate relative to its paired baseline; the connecting line does not imply cumulative gains. The example preserves a simulated quality failure in round 6 and a simulated regression/rollback in round 8. Explicitly recorded cumulative chart specs still use the cumulative view; ambiguous same-round candidates use the independent-results view. Missing evidence in the synthetic campaign remains marked as missing.

视频展示了中文网页从空任务开始随十轮模拟记录实时更新；第 6 轮质量失败、第 8 轮回退。图上的数值均为模拟，不代表真实 GPU 性能。
https://github.com/user-attachments/assets/a366b970-562b-4a90-8dc0-039f7a5b3966

[amy-why-3459]

Review at head debc8d97: not approving yet.

1. Branch-chart fallback can still hang (src/inferrsi/harness/dashboard.py:173-180, loop in src/inferrsi/summary_chart.py:771): the bot's F1 is still unaddressed in code at this head. A finite recorded improvement_pct / target_improvement_pct near float max passes the math.isfinite(spread) guard, tick becomes 1e307, x_max overflows to inf, and while current <= x_max + 1e-9 never terminates, blocking inferrsi dashboard and the single-threaded live server. Expected: bound the synthesized branch spec like _bounded_chain() (return the existing "Recorded values exceed the supported chart scale." note) and make the tick loop stop on a non-finite/non-advancing bound; add regressions for both signs.
2. Unreviewed delta: four commits after 796a5852 (f7f19e11..debc8d97: idle connections, cancelled runs, snapshot compression for remote viewers, trace dedup) have no review yet. In particular, please describe what "remote viewers" means for the live server: is it still loopback-only by default, and is any non-loopback binding opt-in?

[amy-why-3459]

Incremental review debc8d97..3a900852 (one commit, Show recorded candidate and GPU options in dashboard commands), checked against my previous review. Not approving yet.

This delta itself looks reasonable:

• src/inferrsi/harness/dashboard_data.py _safe_argv: --candidate values are shown only when they match a recorded campaign candidate ID, fit [A-Za-z0-9][A-Za-z0-9_.-]{0,127} and do not hit _SENSITIVE_NAME; --gpu only shows 1-3 digit ordinals. load_dashboard passes the known candidate set into _run (L836-837), and record_candidate limits it to the recorded candidate (campaign_records.py:205).
• The new test covers unrecorded IDs, sensitive-looking IDs, long GPU values and IP-like values, and docs/dashboard.md documents the extra visible identifiers.
• CI at 3a900852 is green.

Still open from the previous round (no change to these files in the delta):

1. Branch-chart fallback hang (src/inferrsi/harness/dashboard.py:173-180, tick loop in src/inferrsi/summary_chart.py:771): a finite improvement_pct / target_improvement_pct near float max still passes math.isfinite(spread), so x_max can overflow to inf and while current <= x_max + 1e-9 never terminates, blocking inferrsi dashboard and the single-threaded live server. Expected: bound the synthesized branch spec like _bounded_chain() (return the existing "Recorded values exceed the supported chart scale." note), make the tick loop stop on a non-finite or non-advancing bound, and add regressions for both signs.
2. Remote viewers / binding: still need a short description of what "remote viewers" means for the live server from f7f19e11..debc8d97: is it loopback-only by default, and is any non-loopback bind explicitly opt-in?

Minor question on this delta: --gpu accepts only a single ordinal, so common forms like --gpu 0,1 are redacted. Fine if intended; just confirming.

Static review of the diff and current-head dashboard_data.py; I did not run the tests locally.

[amy-why-3459]

LGTM

PR review verdict — PASS
Reviewed base 2baebf6e and head 9301d29a (delta 3a900852..9301d29a, one commit Bound chart axes for extreme recorded values), checked against my previous review.

Both open points are resolved:

1. Branch-chart fallback hang (bot F1): _chart() now calls the new _bounded_branch() (src/inferrsi/harness/dashboard.py L82-89, L186-187) before building the spec, rejecting non-finite gains, ratios above 1e6 (same bound as _bounded_chain), a non-advancing tick and more than ~150 ticks, and returns the existing "Recorded values exceed the supported chart scale." note. The shared renderer is also hardened: _axis_ticks() (src/inferrsi/summary_chart.py L137-162) rejects non-finite/non-increasing bounds, non-advancing steps and >1000 ticks, and replaces both open-ended tick loops in _build_chain/_build_branch; floor/ceil overflow and non-finite reference gains now raise SpecError, which _chart() already catches. target_gap_pct goes through _number(), so an overflowing gap becomes None instead of a non-finite JSON value. Regressions cover both signs near float max in the offline and live /api/state paths (with build_svg patched to fail if reached), overflowing branch bounds and _axis_ticks edge cases.
2. Remote viewers / binding: docs/dashboard.md now states the server always binds to 127.0.0.1 with no non-loopback option; this matches DashboardServer binding ("127.0.0.1", port) and the CLI exposing only --port. Remote access is via SSH local forwarding.

Ordinary values are unaffected: the synthesized tick keeps spread/tick within roughly 5-50, well under both caps.

All current-head checks pass (Ready full-check 3.10/3.11, installed-package, kv-task-qualification, pr-references, declarations, lint-and-smoke). My earlier minor question about --gpu 0,1 being redacted is still open but not blocking.

Limits: static review of the delta and the affected functions; I did not run the tests locally or open the dashboard in a browser. Agent-submitted under direct delegation from the account owner (@amy-why-3459).

[zengchuang-hw]

@zengchuang-hw review

[zengchuang-hw]

PR review verdict — DO NOT PASS (non-voting)

Following up on the review request. Reviewed base 2baebf6e and head 9301d29a.

1. P2: A candidate with a missing run is presented as recorded evidence. In optimization_story.py:112–120, measured() treats a unique candidate row as a recorded run. Using the existing story fixture, I kept candidate b and moved its associated run directory out of the campaign. load_dashboard() returned state: recorded for both the iteration's measured result and the failure summary's evidence link, even though the outcome and verdict were not_recorded. It emitted no story warning. This can make an incomplete campaign export appear to substantiate a reported failure. Please preserve the distinction between a candidate record and available run evidence, and add a regression that retains the candidate while omitting its run.
2. The expanded scope needs a decision in [Task] RSI Dashboard: Optimization Workflow Visualization and Trace Capture #314. The accepted scope specifies English offline HTML and excludes a service. The PR adds a loopback HTTP service, bilingual UI and optimization story/capture. The current Issue record leaves these additions open. Please link the owner/requester's scope confirmation or split the additions out of this delivery.

The Ready CI and knowledge-policy check passed. On Windows with Python 3.12.1, 183 targeted dashboard, story, trace, campaign and chart tests passed, along with Ruff and the diff whitespace check. The separate missing-run repro above exercises a gap in that suite. Browser behavior and the real-run poor-outcome feedback case remain unverified in this review.

[amy-why-3459]

Incremental review 9301d29a..fd2bbe76 (one commit, Distinguish missing runs from recorded story evidence), head fd2bbe768f2247168bffbbf6922f533b1aca7548, checked against my previous review at 9301d29a (that approval was dismissed by this push) and zengchuang-hw's DO NOT PASS review. Not approving yet.

Resolved in this delta

• Missing run shown as recorded evidence (zengchuang-hw P2 Bootstrap CLI-first RSI experiments and source-grounded diffusion knowledge #1): measured() in src/inferrsi/harness/optimization_story.py (new L118-126) now requires a run_id plus at least one present run object (run_lock/status/evaluation/quality) or a valid events source before returning state: recorded; otherwise it returns not_recorded, or unavailable when a ref is corrupt/unreadable/unsafe_path. resolve() (~L199-218) emits a warning for every non-recorded link and adds outcome_not_recorded for a recorded run whose outcome is missing. The five new tests in tests/test_optimization_story.py cover exactly the requested repro (candidate kept, run directory moved away), all-corrupt run files, a running candidate, a recorded failure with partially corrupt files, and an event-only run. The new strings are wired into the translation table and measuredText() in template.html, with assertions in tests/test_dashboard_translation_ui.py; docs/dashboard.md documents the new check.

Still open before I can approve

1. CI at fd2bbe76 is not finished: full-check (3.10) and full-check (3.11) were still in progress when I looked (lint-and-smoke, installed-package, declarations, pr-references, kv-task-qualification passed). I'll re-check once they complete.
2. Scope confirmation for [Task] RSI Dashboard: Optimization Workflow Visualization and Trace Capture #314 (zengchuang-hw P2 feat: bootstrap CLI experiment workbench and diffusion knowledge #2): the scope amendment is posted from the PR author's account and says it records the requester's direction on their behalf. Could the requester confirm it directly in [Task] RSI Dashboard: Optimization Workflow Visualization and Trace Capture #314, or could you link their original instruction? I can't verify the expanded scope (live loopback service, bilingual UI, story/capture) from the record as it stands.

Question (non-blocking): in the iteration view, a candidate whose run exists but has no outcome (the running case) stays state: recorded with outcome: not_recorded and gets no iteration-level warning, while the same candidate cited in a summary becomes outcome_not_recorded with a warning. Is that asymmetry intended? If so, a one-line note in docs/dashboard.md would help.

[amy-why-3459]

Incremental review fd2bbe76..10f1916d (one commit, Clarify recorded runs without outcomes), head 10f1916d1fd93623b8ccbd88b808b068cc1c1271, checked against my previous review at fd2bbe76. Not approving yet.

Resolved / cleared in this pass

• Docs note for the running-vs-summary asymmetry (non-blocking question): docs/dashboard.md now states that a recorded iteration means available run evidence, not a completed result, and that a success/failure summary citing a run with no recorded outcome is flagged. That matches the measured() / resolve() behavior left at fd2bbe76 (state: recorded + outcome: not_recorded in the iteration view; outcome_not_recorded + warning when cited from a summary). No code files changed in this delta.
• CI at fd2bbe76: Ready checks completed successfully after my previous comment (full-check 3.10/3.11, installed-package, lint-and-smoke, declarations, pr-references, kv-task-qualification).

Still open before I can approve

1. Scope confirmation for [Task] RSI Dashboard: Optimization Workflow Visualization and Trace Capture #314 (zengchuang-hw P2 feat: bootstrap CLI experiment workbench and diffusion knowledge #2 / my prior blocker): the scope amendment is still posted from the PR author's account and says it records the requester's direction on their behalf. I still do not see a direct requester confirmation or a link to the original instruction in [Task] RSI Dashboard: Optimization Workflow Visualization and Trace Capture #314. Please have the requester confirm the expanded scope (live loopback service, bilingual UI, story/capture) directly, or link their original instruction.
2. CI at current head 10f1916d: when I looked, full-check (3.10) and full-check (3.11) were still in progress (lint-and-smoke, installed-package, declarations, pr-references, kv-task-qualification already green). Docs-only delta, but I'll re-check once the Ready full checks finish.

No new code defect found in this docs-only delta. Static review of the commit and the surrounding dashboard.md section; I did not run the tests locally or open the dashboard in a browser. Agent-submitted under direct delegation from the account owner (@amy-why-3459).

[zengchuang-hw]

P3 · Keep summary links to runs that have evidence and no recorded outcome

At 10f1916d, a success or failure summary citing a running candidate correctly displays “No recorded outcome.” The new outcome_not_recorded state retains the candidate’s run ID, but the summary renderer disables the button for every state except recorded. The candidate details remain reachable from the iteration view.

I reproduced this with the running-candidate fixture, load_dashboard(), and the actual summary renderer in a Node DOM stub. Please keep the missing-outcome label and warning and let the summary link open the available candidate details. A renderer assertion for the enabled button would cover this path.

[amy-why-3459]

Incremental review 10f1916d..2218beab (one commit, Keep summary navigation for runs without outcomes), head 2218beabfea8e89f4c221dc17a40f1d8e9353ada, checked against my previous review at 10f1916d. Not approving yet.

Resolved in this delta

• zengchuang-hw P3 (summary link disabled for runs with evidence but no outcome): src/inferrsi/data/dashboard/template.html L825 now enables the summary chip for recorded and outcome_not_recorded and keeps it disabled for every other state; the No recorded outcome label from measuredText() is unchanged. Non-candidate refs (else branch) are untouched.
• Regression: tests/test_optimization_story.py::test_summary_navigates_to_run_without_outcome_but_not_missing_runs builds the running-candidate fixture, asserts state: outcome_not_recorded + run_id: run-b + the has no recorded outcome check, then executes the real selectCandidate / renderSummary source in a Node stub: the b button is enabled and selects b, while not_recorded / unavailable / ambiguous buttons stay disabled and do not change the selection. That covers the requested renderer assertion.
• CI at 10f1916d and 2218beab: all checks are green at the current head (full-check 3.10/3.11, installed-package 3.10/3.11, lint-and-smoke, declarations, pr-references, kv-task-qualification), which clears my previous item 2.

Question (non-blocking)

• The new test is skipif(shutil.which("node") is None). Could you confirm full-check actually runs it rather than skipping it (e.g. the runner has Node on PATH, or the job log shows it as passed)? Otherwise this UI path is only covered locally.

Still open before I can approve

1. Scope confirmation for [Task] RSI Dashboard: Optimization Workflow Visualization and Trace Capture #314: no new comment in [Task] RSI Dashboard: Optimization Workflow Visualization and Trace Capture #314 since the scope amendment, which is still posted from the PR author's account on the requester's behalf. Please have the requester confirm the expanded scope (live loopback service, bilingual UI, story/capture) directly in [Task] RSI Dashboard: Optimization Workflow Visualization and Trace Capture #314, or link their original instruction.

No new code defect found in this delta. Static review of the commit and the surrounding renderer/test code; I did not run the tests locally. Agent-submitted under direct delegation from the account owner (@amy-why-3459).

[zengchuang-hw]

Following up on the CI coverage question. On the same main base (9dded7ad), the prior Ready full-check jobs on 10f1916d each reported 3,695 passed, 13 skipped (3.10, 3.11). At 2218bea, both report 3,696 passed, 13 skipped (3.10, 3.11). The delta adds exactly one test, the summary-navigation Node regression. The unchanged skip count and extra pass show it ran on both versions, although the default pytest output reports aggregate counts. The local exact-head run also passed all 36 targeted tests with no skips.

[amy-why-3459]

LGTM

PR review verdict — PASS
Reviewed base 2baebf6ebb815ebf07d21c41e3b3754858c0798e and head 2218beabfea8e89f4c221dc17a40f1d8e9353ada.
Incremental re-review after my prior approval on 9301d29a was dismissed by new pushes (9301d29a..2218beab, 3 commits).

Verified:

• optimization_story.measured() now distinguishes not_recorded / unavailable / ambiguous / recorded, and success/failure links can surface outcome_not_recorded when run evidence exists but no outcome is frozen.
• Summary chips stay navigable for recorded and outcome_not_recorded, and stay disabled for missing / unavailable / ambiguous evidence (template.html).
• Docs and EN/ZH translation UI cover the new states; added regressions in tests/test_optimization_story.py and tests/test_dashboard_translation_ui.py match the intended behavior.
• Current-head Ready checks are green (full-check 3.10/3.11, installed-package, kv-task-qualification, pr-references, declarations, lint-and-smoke).

Static review of the incremental delta and hosted checks only; I did not run the Node UI harness or open a live dashboard.

Issue #314 scope/acceptance ownership remains separate from this software review.

[zengchuang-hw]

@zengchuang-hw review

[zengchuang-hw]

Recheck 2218beab..2218beab for the new review request. Current head remains 2218beabfea8e89f4c221dc17a40f1d8e9353ada. The prior technical follow-up applies to the same content. No additional implementation blocker was found; scope provenance remains pending.

本轮已解决

The missing-run and summary-navigation findings remain fixed at this head. The Ready CI and knowledge status passed. Amy's new approval is bound to this head and explicitly separates software review from Issue scope ownership.

批准前必做（Before I can approve）

• Existing owner david6666666 needs to supply direct requester confirmation of the bilingual UI, optional loopback live serving and caller-authored story/context amendment, or link the original authorization. That closes the existing scope-provenance item. The current #314 status still records it as pending, and its amendment remains author-posted.

非阻塞问题或建议

No additional finding. The poor-outcome feedback demonstration remains separate Issue acceptance work.

Checks and limits

This recheck refreshed the head, checks, reviews and scope discussion. It reuses the same-head validation recorded in the prior follow-up, including 36 local targeted passes and the Node renderer regression. No new local or live-browser run was performed.

Codex performed this non-voting recheck through zengchuang-hw, under the account holder's direct review delegation.