Repository navigation
Incentive monitor v2: status/LULD, corporate-action events, screeners, 20 s sweeps, option IV, FINRA short volume (board v2) - #220
Merged
Conversation
…and a preregistered paper forward study Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…t order Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…, exclusive ledger, operating companies, exact analysis) Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…isions within 90 minutes of an early close Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…cans, full Trust disclosure, retry rule Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…separate from paper performance Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…reener, option chains, FINRA short volume, budgeted sweeps, board v2 Sources added to monitor.py, each recorded with receive times in owner-only files and held to an explicit per-source call budget written to monitor.jsonl: - IEX stock stream (v2/iex; v2/sip stays with the engine): statuses, LULD bands and imbalances for every symbol, merged with the Nasdaq RSS into one halt state (latest event by its own time wins); message counts per type and tape measure tape coverage (CTA statuses on IEX are unobserved so far). A host lease per stream endpoint refuses a second holder before it connects; a 406 on the IEX endpoint waits instead of contending. - Corporate-action events (SSE on the documented stream host): every event archived, deduplicated by ULID event id, resumed with Last-Event-Id; splits, mergers, spin-offs and name/symbol changes surfaced as board context from a bounded 30-day replay plus the live stream. - Screener: most actives by volume and by trades, top movers (3 calls per sweep). - Option chain snapshots for the top 25 candidates by the v1 score (one page each), near-the-money implied volatility and a per-root IV change against the previous session (same expiry) or the first value today; open interest from the contracts endpoint once per root per day. - FINRA daily consolidated short sale volume: at most one request per sweep, after 18:05 ET, stored once with a manifest; header and trailer count checked. - Sweep cadence 20 s by default; the plan is refused (exit 2) above 5% of the 10,000/min data limit, above 5% of the 200/min trading limit, or with the Nasdaq RSS polled more than once a minute; a call over its source cap is refused before any request. Board v2 (board-v2.json, board_version 2) keeps every v1 component and weight (score_v1) and adds fixed, unvalidated components: stream_news_halt 2.0, stream_volatility_halt 1.0, corporate_action 0.5, iv_runup 1.0, short_volume_high 0.5. board.json stays board version 1. board_scan.py refuses any board whose board_version differs from the protocol's (forward protocol v1 pins 1; the frozen protocol file is unchanged). Evidence: synthetic fixtures and one local-integration once-sweep against a synthetic response router (no network); docs.alpaca.markets and FINRA pages re-read 2026-09-24; the monitor was not run against live endpoints. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…review fixes Resolve the two independent reviews of 736c555. - Coexistence: every run declares --standalone or --follow-v1 V1_OUT. A follower holds no news/OPRA stream, polls neither EDGAR nor the Nasdaq RSS (zero budgets), reads version 1's appended lines and writes board-v2.json only. --standalone refuses while a version 1 monitor runs on this host (/proc). Leases live in one fixed per-user directory. The stream backoff resets only after data or 60 s connected, and 402/405/409/410 stop a stream. - Options: chains and IV sampled 09:45-16:00 ET only. The previous-session baseline is the last in-session row, the first-today baseline is per expiry, and IV older than two sweeps no longer counts. The expiry window is 35 days (it always holds a monthly), an empty in-band chain is retried once without the band, and a chain poll has a deadline. Open interest is opt-in and paused near the trading limit. - Boards: one scoring pass feeds both boards, and board.json is written before any v2 work, which is guarded (board_v2_error). Halts keep v1 precedence across sources, and any RSS halt today blocks stream_volatility_halt. - Corporate actions: US region only, first-insert time kept, and malformed symbols ignored. The archive resumes from its last id whatever its age and records ca_resume_gap only when that id is not redelivered. An empty archive seeds from the retained history; the context replay is bounded and runs on the stream thread. - FINRA: every failure waits an hour. Each stored file is checked once more after the next publication, and superseded versions are kept. - Budgets: the daily-bar load resumes at the failed request under a per-minute cap, and --rss-seconds 0 is refused. - Halts: a multi-day halt enters the merged state, and a moved resumption replaces the earlier one. - Frozen v1: the monitor refuses to run beside protocol v1's frozen bridge, and board_scan refuses protocol v1 decisions under non-frozen code (code_not_frozen). Evidence (synthetic fixtures / local integration, no network): 121 tests OK, 0 skipped, under python3 and the pinned engine venv; scripts/validate.py passed. Every new test fails against 736c555, and 54 fix-revert mutants plus the builder's 10 are all killed. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
|
You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard. |
Owner
Author
|
CI note (2026-09-25). Nothing in the PR needs to change. The next push, or a close and reopen, runs the current 🤖 Generated with Claude Code |
seathatflowsinourveins
added
the
lane:trading
This was referenced Sep 25, 2026
Owner
Author
|
Trading lane, 2026-09-25: carrying this PR to merge, after #187. The failing Plan, following the dormant-owner merge etiquette:
The original owner can take this back by replying here. 🤖 Generated with Claude Code |
… merge #187 was squash-merged as 3e4054d, so its files were added on both sides. The five files this PR changes conflicted and resolve to this PR's head (1f136ba) byte for byte. The merged tree differs from main by exactly this PR's own delta over #187 (05c2849..1f136ba; 5 files, +3529/-122). Main's monitor.py and board_scan.py still match FROZEN_V1. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
…1 as a v2 input board_scan.ENGINE_FILES hashes 8 engine files. The transitive import closure of mover_runner.py within adaptive-paper/ adds 14 more modules that the list omits, measured on main after #187: - 4 existed when v1 froze (05c2849) and have changed since; - 6 are unchanged since the freeze; - 4 were added after it, including corporate_actions and financing from the engine P0 release. The v1 bridge is frozen, so the gap stays in v1 as a recorded limitation. Protocol v2 should bind the whole engine (every *.py, or a clean-checkout tree hash) plus the runtime lock before its pre-outcome review. No v1 order has been placed. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
![github-advanced-security[bot]](https://avatars.githubusercontent.com/in/57789?s=60&v=4){kind=small}
…tname (CodeQL) CodeQL flagged py/incomplete-url-substring-sanitization at four places in tests/test_incentive_monitor.py: "nasdaqtrader.com" in url and "sec.gov" in u. A substring test also matches a lookalike host or a query string. A host_is() helper now compares urlsplit(url).hostname with the domain and its subdomains. The router, both URL-count assertions and the unflagged sec.gov assertion use it. host_is rejects ?u=nasdaqtrader.com, nasdaqtrader.com.evil.example and notsec.gov. 121 tests pass on system Python and on the pinned runtime. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Owner
Author
|
Merged as f625b07 (trading lane, 2026-09-25). How it was carried after #187's squash merge (3e4054d):
Added commits:
Checks:
No v1 or v2 order has been placed. The paper keys currently return HTTP 401. 🤖 Generated with Claude Code |
seathatflowsinourveins
added a commit
that referenced
this pull request
Sep 25, 2026
… incentive board and data lake, and the recounted WSL host (#288) - simulation: 866 to 884 fills per min at 70 ms exact release, 30 of 30 minutes, and 672 at 1 s (#254, sim-capacity/receipts/20260925-elite-tier.json). - paper: adds the margin financing model (#243); still blocked on the paper keys (HTTP 401). - mover-early-entry-studies: v1 to v3, with the v3 core merged (#190) and not frozen. - New gates: - incentive-board-forward (#187, #220): 0 orders; - data-lake-source (#208): minute-bar pilot not run. - wsl-workstation-host: the foundation lane's wording. Counts were recounted on main 0074a0c by the last independent_session verdict per receipt, superseded excluded: 69 receipts, 56 agree, 1 needs changes, 0 awaiting review, 12 superseded (#285). The foundation lane acked the strings and the recount rule. Co-authored-by: Scout <scout@local> Co-authored-by: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
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
2 participants
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 (stacked on #187)
Monitor v2 uses more of the entitled and free data feeds for incentive finding. Each source records receive timestamps, and each call budget is written to
monitor.jsonl.New sources:
Board v2:
board_versionfield is added.board_scanrefuses a board whose version differs from its protocol's, so the running forward study (v1) stays on v1.Review trail
Two independent reviews found 5 majors, all fixed in 1f136ba. The recheck approved, with 6 minors left.
Checks
tests.test_incentive_monitor tests.test_incentive_forward: 121 OK. Synthetic fixtures only; no live run.python3 -m unittest: 4,065 OK (419 skipped).scripts/validate.py: passed.Next
Forward protocol v2 pairs board v2 with the engine P0 release. It gets its own pre-outcome review before any v2 order.
🤖 Generated with Claude Code