Repository navigation
Native faults: the 2026-10-01 run on engine dca821cc becomes the gate's current receipt (native_faults_passed) - #584
Merged
Conversation
…e gate's current receipt receipt.json is now the 2026-10-01 14:38:32Z Alpaca paper run of the frozen native-faults plan (C01, C02, C05, C04) on engine dca821c, status native_faults_passed, exit 0, with the same four outcomes as the earlier passing runs. It binds dca821c's runner.py, safety.py and transport.py, so scripts/trading_gates.py reports the native-fault-behaviour binding as bound 6, stale 0 (it was stale 3). - Retain the 2026-09-25 receipt byte-for-byte as receipt-20260925t182513.json and point its run record's subject_receipt there. - Add the sanitized run record evidence/run-20261001t143832.json and the independent observation evidence/independent-observation-20261001t143832.json, with its stdout byte-identical. The observation uses a changed method: the coordinator's alpaca-py TradingClient GETs with the engine's guarded credential loader. - README: a new top section with the results table and what this run did not check (no host-clock measurement, no separate account-identity probe), and updates to the sentences that named the current receipt or engine. - Gate note: prepend the dated 2026-10-01 rebinding; status unchanged. Co-Authored-By: Claude Opus 5.5 <[email protected]>
This is the hot-file commit (docs/lanes.md protocol), made on main 758bce0's manifest. It registers four new native-faults files and refreshes the hashes of README.md, receipt.json, the 2026-09-25 run record and gates-20260922.json. component_matrix.py --write and new_host_grand_list.py --write left their reports unchanged. Co-Authored-By: Claude Opus 5.5 <[email protected]>
…ipt; runtime-target rebinding sentence for 2026-10-01 README-safety.md and README-transport.md named native-faults/receipt.json as the 2026-09-25 run; that receipt is now receipt-20260925t182513.json and receipt.json is the 2026-10-01 run on engine dca821c. catalogs/us-equities/runtime-target.json's native-fault sentence gains a dated 2026-10-01 rebinding (engine_sources_sha256 runner.py 34ab492f..., safety.py f43def73..., transport.py 8d1222f7...), keeping the 2026-09-25 rebinding as history. Co-Authored-By: Claude Opus 5.5 <[email protected]>
…ocol) manifests/evidence.json is main's copy (d6d3bf7) with this branch's 11 files re-registered (the receipts, run record, observation and observer output, README, the two engine READMEs, gates-20260922.json and runtime-target.json); no receipts[] or convergence_records[] entry. component_matrix.py --write and new_host_grand_list.py --write re-run. scripts/validate.py exit 0; scripts/evidence_manifest.py --check passed (8905 files); validate_catalogs exit 0; tests.test_native_faults_min, test_trading_gates, test_trading_gates_bindings, test_adaptive_paper_source_hashes_manifest and test_catalogs Ran 116, OK (skipped=1); scripts/trading_gates.py: native-fault-behaviour bound 6, stale 0. Co-Authored-By: Claude Opus 5.5 <[email protected]>
|
You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard. |
…e observer script Independent Opus review of 0020911: approve, with 4 low documentation findings, all repaired: - runtime-target.json: the closing sentence now says the qualification covered the 2026-09-25 engine and since 2026-10-01 covers dca821c, and names the three added limits. - README: "fresh state root" now says why (inside the run's output directory, which the run script refused to reuse); the no-ladder condition states what ran on the host (no order-throughput ladder session; account 1's mover series with its own writer lock). - README: two limits added (ambiguous or in-flight submission faults not exercised; the observer's --after value and interpreter come from the coordinator's command); a sentence explains why the 2026-09-25 observer stdout still names native-faults/receipt.json. - Gap closed: the observer script is committed byte-identical as evidence/observe-native-faults-20261001.py (sha256 f118798f...); the observation record records its --after value, exit code 0 and runtime; the run record describes the harness stderr (no identifier). Outside the native-faults folder: the owner's verdict record no longer calls dca821c's native-fault binding stale. Co-Authored-By: Claude Opus 5.5 <[email protected]>
…e protocol, last commit) manifests/evidence.json re-registers the README, the observation and run records, the committed observer script, runtime-target.json and observability/grand-dashboard/state.json through scripts/host_receipts.register_file. state.json: the alpaca-paper-operational gate row now reads "native faults re-verified on fee-aware engine 2026-10-01". component_matrix and new_host_grand_list re-run. scripts/validate.py exit 0; evidence_manifest --check passed (8906 files); tests.test_trading_gates, test_trading_gates_bindings, test_catalogs and test_grand_dashboard Ran 102, OK; trading_gates.py native-fault-behaviour bound 6, stale 0. Co-Authored-By: Claude Opus 5.5 <[email protected]>
Collaborator
Author
Independent review (Claude Opus, read-only, at
|
| # | Finding | Repair |
|---|---|---|
| 1 | runtime-target.json's closing sentence still said the qualification covered "the released engine". |
It now says the qualification covered the 2026-09-25 engine and since 2026-10-01 covers dca821cc, and it names the three added limits. |
| 2 | README: "fresh state root" was not supported by the run record, and the host-wide no-ladder condition was stated only for account 2. | The README now says why the state root was new: it lies inside the run's output directory, which the run script refused to reuse. It also states what ran on the host: no order-throughput ladder session, and account 1's mover series with its own writer lock. |
| 3 | Two limits were in the run record but not in the README: ambiguous or in-flight submission faults were not exercised, and the observer's inputs were not recorded. | Both added. |
| 4 | The retained 2026-09-25 observer stdout still names native-faults/receipt.json. |
It stays a byte-identical copy; a README sentence explains that its receipt sha256 (31c112e3…) identifies today's receipt-20260925t182513.json. |
| gap | The observer script was not committed. | It is committed byte-identical as evidence/observe-native-faults-20261001.py (sha256 f118798f…). The observation record now carries its --after value, exit code 0 and runtime. |
| gap | The harness stderr was not described. | The run record describes it: two notices from the bash -ic wrapper and one alpaca-py stream reconnect notice, with no identifier. |
Also in this round, both outside the native-faults folder:
- The owner's verdict record no longer calls
dca821cc's native-fault binding stale. - The grand dashboard's
alpaca-paper-operationalgate row now reads "native faults re-verified on fee-aware engine 2026-10-01". It is registered in the last commit.
Checks at 0094c3ac:
| Command | Exit | Result |
|---|---|---|
scripts/validate.py |
0 | passed |
scripts/evidence_manifest.py --check |
0 | 8906 files |
tests.test_trading_gates, test_trading_gates_bindings, test_catalogs, test_grand_dashboard |
0 | Ran 102, OK |
scripts/trading_gates.py |
0 | native-fault-behaviour bound 6, stale 0 |
One review round and one repair round.
Merged
4 tasks done
seathatflowsinourveins
added a commit
that referenced
this pull request
Oct 1, 2026
…ion-broker row after #584 (#586) PR #584 (merged as 752714a) made the 2026-10-01 native-fault run on engine dca821c the gate's current receipt. The row's notes said that dca821c had no paper execution and that its native-fault binding was stale; the owner replaced that sentence verbatim. The edition file is not a registered manifest row, so no hot-file commit follows. Co-authored-by: Scout <scout@local> Co-authored-by: Claude Opus 5.5 <[email protected]>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What
The
native-fault-behaviourgate's receipt is rebound to the merged fee-fix enginedca821cc(#559). The fee fix changedrunner.py,safety.pyandtransport.py, soscripts/trading_gates.pyhad reported the gate's source binding stale since then (report-only). That binding is now current.The run.
harness.pysha25619d0a9b9…,plan.jsonf276b26c…).dca821cc, with the pinned runtime (Python 3.12.3, alpaca-py 0.44.0, nautilus_trader 2.0.0rc5).Result:
native_faults_passed.Two posts were reserved, and cleanup proved the account flat. An independent read-only observation afterwards (alpaca-py GETs, outside the harness) saw:
c01, SPY buy 1 limit 380.73 day, canceled, 0 filled;c04client-id lookup: 404;Files
native-faults/receipt.json: the new receipt, unchanged from the harness (sha25654430f38…). The 2026-09-25 receipt is retained byte-for-byte asreceipt-20260925t182513.json.native-faults/evidence/run-20261001t143832.json: the run record in the 2026-09-25 schema. It holds:native-faults/evidence/independent-observation-20261001t143832.jsonand the observer's stdout. The record states plainly that the observation method differs from 2026-09-25: alpaca-pyTradingClientGETs through the engine's guarded credential loader, not a standard-library observer.native-faults/README.md: a new top section with the results and what this run did not check. Every sentence that named the current receipt or engine is updated.catalogs/us-equities/gates-20260922.json: a dated rebinding sentence in the gate note.statusstaysestablished.catalogs/us-equities/runtime-target.json,adaptive-paper/README-safety.md,README-transport.md: their references to the current receipt.manifests/evidence.json: 11 files re-registered in the last (merge) commit; noreceipts[]orconvergence_records[]entry.Limits (also in the records)
This run is weaker than the 2026-09-25 run in three ways:
chronycis not installed on this host.Checks
tests.test_native_faults_min,test_trading_gates,test_trading_gates_bindings,test_adaptive_paper_source_hashes_manifest,test_catalogstest_native_faults_min+test_adaptive_paper_source_hashes_manifest(before the run, from the clone)scripts/trading_gates.pynative-fault-behaviourbound 6, stale 0 (was stale 3)scripts/validate.pyscripts/evidence_manifest.py --checkscripts/validate_catalogs.pySOTA sources
blueprints/us-equities/adaptive-paper/native-faults/harness.py(19d0a9b9…) andplan.json(f276b26c…), unchanged since the 2026-09-25 run.dca821cca85dce3647fa7b488d5a23fbe5b85d4a(Adaptive paper: book Alpaca FEE activities in ledger cash accounting (unblocks account 2's IOVA recovery) #559). Runtime pins: alpaca-py 0.44.0 and nautilus-trader 2.0.0rc5 (blueprints/us-equities/adaptive-paper/requirements.txt).docs/acceptance-evidence-policy.md. Hot-file protocol:docs/lanes.md.🤖 Generated with Claude Code