dexbot 1.4.19 → 1.4.21

package/docs/COPY_ON_WRITE_MASTER_PLAN.md

@@ -257,7 +257,7 @@ Dependency utilities merged into `modules/order/utils/order.ts` during v0.6.0-pa
 ### Build Step 2: Core Integration ✅
 - `performSafeRebalance()` → delegates to `_applySafeRebalanceCOW()`.
 - `_applySafeRebalanceCOW()` — creates working grid, runs planning, returns result without modifying master.
-- `_reconcileGridCOW()` — delta reconciliation against working copy.
+- `WorkingGrid.buildDelta()` — delta reconciliation against working copy (`modules/order/working_grid.ts:204`, delegating to `utils/order.ts`).
 - `_commitWorkingGrid()` — atomic swap from working to master.
 
 ### Build Step 3: Broadcast Integration ✅
@@ -362,7 +362,7 @@ await updateOrdersOnChainBatch(cowResult); // Execute via COW
 |--------|-------------|
 | `performSafeRebalance(fills, excludeIds)` | Entry point — delegates to COW |
 | `_applySafeRebalanceCOW(fills, excludeIds)` | Creates working grid, runs planning |
-| `_reconcileGridCOW(targetGrid, boundary, workingGrid)` | Delta against working copy |
+| `WorkingGrid.buildDelta(masterGrid)` | Delta between master and working copy (`modules/order/working_grid.ts:204`, delegating to `utils/order.ts`) |
 | `_commitWorkingGrid(workingGrid, indexes, boundary, options = {})` | Atomic swap to master |
 | `_setRebalanceState(state)` | Track rebalance state |
 | `_currentWorkingGrid` | Reference to working grid during rebalance for fill sync |
@@ -479,27 +479,18 @@ Stale-Plan & Stack Discipline Tests (v1.4.8):
 ## Appendix C: Constants Added (`modules/constants.ts`)
 
 ### COW Performance Thresholds
-- `COW_PERFORMANCE.MAX_REBALANCE_PLANNING_MS` — max time for rebalance planning phase.
-- `COW_PERFORMANCE.MAX_COMMIT_MS` — max time for grid commit operation.
-- `COW_PERFORMANCE.MAX_MEMORY_MB` — memory threshold for working grid operations.
-- `COW_PERFORMANCE.INDEX_REBUILD_THRESHOLD` — grid size threshold for index rebuilding.
+- `COW_PERFORMANCE.MAX_REBALANCE_PLANNING_MS` — planning-phase duration above which a slow-plan warning is logged (100ms).
+- `COW_PERFORMANCE.GRID_MEMORY_WARNING` — working grid size (bytes) that triggers a memory warning (5,000).
 - `COW_PERFORMANCE.WORKING_GRID_BYTES_PER_ORDER` — estimated memory per order (500 bytes).
+- `COW_PERFORMANCE.MAX_OPS_PER_BROADCAST` — maximum order operations (creates, updates, cancels) per broadcast transaction (4); larger batches are split into sequential broadcasts.
 
 ### Pipeline Timing
-- `PIPELINE_TIMING.MAX_FEE_EVENT_CACHE_SIZE` — LRU cache limit for fee dedup (10,000 entries).
-- `PIPELINE_TIMING.FEE_EVENT_DEDUP_TTL_MS` — fee event deduplication TTL (6 hours).
-- `PIPELINE_TIMING.CACHE_EVICTION_RETENTION_RATIO` — LRU eviction retention (0.75).
 - `PIPELINE_TIMING.RECOVERY_DECAY_FALLBACK_MS` — recovery decay fallback (180 seconds).
 
 ### Grid & Timing
 - `GRID_LIMITS.RELATIVE_ORDER_UPDATE_THRESHOLD_PERCENT` — relative threshold (%) for in-memory COW order equality checks.
-- `GRID_LIMITS.STATE_CHANGE_HISTORY_MAX` — max state change history entries (100).
 - `TIMING.LOCK_REFRESH_MIN_MS` — minimum lock refresh interval (250ms).
 
-### Fee Dedup Precision
-- Fee-event dedupe keys now quantize size with `floatToBlockchainInt(size, orderPrecision)` (derived from BUY/SELL side precision).
-- Fixed `1e8` satoshi conversion is no longer used.
-
 ## Appendix D: Validation Gates
 
 Run these tests before promotion:

package/docs/COW_INVARIANTS.md

@@ -211,7 +211,7 @@ This document defines the non-negotiable behavioral invariants for the DEXBot2 s
 
 - `INV-BATCH-001` Illegal state batch abort
   - `executeBatch` throws `ILLEGAL_SPREAD_STATE` on an illegal grid layout (emitted at `modules/order/utils/validate.ts`, propagated via `modules/order/manager.ts` `_throwOnIllegalState`).
-  - The `_handleBatchHardAbort` catch for `ILLEGAL_ORDER_STATE` (`dexbot_state_recovery.ts:126`) is a test-only dead branch — production never emits that code; only the test stub at `tests/test_patch17_invariants.ts:396` uses it.
+  - The `_handleBatchHardAbort` catch for `ILLEGAL_ORDER_STATE` (`dexbot_state_recovery.ts:133`) is a test-only dead branch — production never emits that code; only the test stub at `tests/test_patch17_invariants.ts:396` uses it.
   - In production, recovery + cooldown are armed on the next maintenance tick via `_abortFlowIfIllegalState` (the `INV-MAINT-002` path), returning `abortedForIllegalState: true` to the caller. The caller does not need to return immediately; the maintenance tick handles recovery.
   - Hard abort triggers one immediate recovery sync (`_triggerStateRecoverySync`) plus arms one maintenance cooldown cycle (`_maintenanceCooldownCycles = Math.max(current, 1)`).
 
@@ -230,7 +230,7 @@ This document defines the non-negotiable behavioral invariants for the DEXBot2 s
     - NOT virtualize the slot.
     - Preserve `orderId` until sync reconciles it.
     - NOT mark the order as stale-cleaned.
-  - Fast path: if the batch result indicates `ORDER_SIZE_DRIFT_TARGETED` (`dexbot_state_recovery.ts:247`), a targeted repair applies the correction directly and skips `_triggerStateRecoverySync`.
+  - Fast path: if the batch result indicates `ORDER_SIZE_DRIFT_TARGETED` (`dexbot_state_recovery.ts:263`), a targeted repair applies the correction directly and skips `_triggerStateRecoverySync`.
 
 ---
 
@@ -254,7 +254,7 @@ This document defines the non-negotiable behavioral invariants for the DEXBot2 s
 - `INV-REG-001` Cross-bot allocation ≤ proportional share
   - Per-bot committed amounts (sum of on-chain orders) must not exceed `totalChainBalance × allocatedPercent`.
   - Violation triggers an error-level log entry (not silent), with tolerance `max(PERCENT_TOLERANCE * 3, 0.15)`.
-  - Registry registration is pre-flight + atomic; only shared-account bots register (`dexbot.ts:521` filters `accountGroups[a].length > 1`), and registration completes before any shared-account bot starts.
+  - Registry registration is pre-flight + atomic; only shared-account bots register (`dexbot.ts:535` filters `accountGroups[a].length > 1`), and registration completes before any shared-account bot starts.
   - Release happens in `DEXBot.shutdown`.
 
 - `INV-REG-002` Async-locked registry writes
@@ -290,7 +290,7 @@ This document defines the non-negotiable behavioral invariants for the DEXBot2 s
 
 - `INV-BROADCAST-002` Deadlock-free reconcile after uncertain broadcast
   - `_reconcileAfterUncertainBroadcast` does not need a `fillLockAlreadyHeld` flag because `AsyncLock` is re-entrant — a second `acquire()` from within the same execution context runs the callback directly instead of queueing.
-  - The `gridLockAlreadyHeld` flag has been eliminated — the lock hierarchy was corrected so that `_syncLock(2)` is acquired before `_gridLock(3)` in all paths.
+  - The lock hierarchy requires `_syncLock(2)` to be acquired before `_gridLock(3)` in all paths; no `gridLockAlreadyHeld` bypass flag exists.
 
 - `INV-BROADCAST-003` Verify-before-retry; never re-sign an uncertain broadcast
   - An uncertain outcome (RPC timeout, connection dropped with a response pending, unknown code) must never be re-signed — a re-sign would land a duplicate transaction on chain (new tx ID per signature).

package/docs/CREDENTIAL_SECURITY.md

@@ -288,16 +288,6 @@ The failure path is consistent regardless of whether the daemon or the
 interactive password prompt handled the authentication, making the output
 predictable for monitoring and alerting.
 
-### Legacy vault migration
-
-Older vaults stored a plain **SHA-256** hash of the master password for
-verification. This hash is deliberately weak by modern standards. The v2
-scrypt-derived vault format replaced it entirely: `masterPasswordHash` is no
-longer read, written, or migrated. A legacy vault that has not been converted
-is rejected by `unlockWithPassword` with "Unsupported key vault format" —
-recreate `profiles/keys.json` with the current key manager to adopt the v2
-format (HMAC-SHA256 vault verifier, scrypt-derived key).
-
 ---
 
 ## 6. Startup Path — Daemon-First, Interactive Fallback
@@ -391,7 +381,6 @@ accounts or after an unclean exit.
 | `probeBootstrapSocket` (live probe before cleanup) | Bootstrap dir cleanup | Prevents removing a live bootstrap directory |
 | `delete process.env.DEXBOT_CRED_BOOTSTRAP_PATH_FILE` | Daemon startup | Bootstrap path cannot be inherited by child processes or read from /proc |
 | Attempt limit (3) + immediate exit | Interactive auth | Limits online brute-force window |
-| `masterPasswordHash` removed | Legacy vault upgrade | Weak SHA-256 verifier no longer read, written, or migrated |
 
 ---
 

package/docs/DEXBOT_COMPARISON.md

@@ -1,7 +1,7 @@
 # DEXBot vs DEXBot2 — Detailed Comparison Report
 
 > **Date:** 2026-08-19 *(metrics refreshed against local source trees)*
-> **Scope:** Full architectural, functional, and operational comparison between the original [DEXBot](https://github.com/Codaone/DEXBot) (Python, v1.0.0) and DEXBot2 (TypeScript, v1.4.19).
+> **Scope:** Full architectural, functional, and operational comparison between the original [DEXBot](https://github.com/Codaone/DEXBot) (Python, v1.0.0) and DEXBot2 (TypeScript, v1.4.21).
 > **Audience:** Developers, contributors, and operators evaluating or migrating between the two projects.
 
 ---
@@ -36,7 +36,7 @@
 
 | Attribute | DEXBot (original) | DEXBot2 |
 |---|---|---|
-| **Release Track** | 1.0.0 | v1.4.19 |
+| **Release Track** | 1.0.0 | v1.4.21 |
 | **Language** | Python 3.6+ | TypeScript 5.x |
 | **Status** | Released 1.0.0, unmaintained | Active development |
 | **Last Repo Activity** | May 23, 2020 | 2026-08-19 |
@@ -46,7 +46,7 @@
 | **Target Exchange** | BitShares DEX | BitShares DEX |
 | **Lines of Code** | ~10,846 Python LOC in `dexbot/` | Large TypeScript codebase; core runtime, adapter, analysis, Claw, and test modules |
 | **Source Files** | 72 Python files in `dexbot/` | ~527 TS files across the repo |
-| **Test Files** | 16 Python test files | 249 `test_*.ts` files (249 auto-discovered via `globSync`) |
+| **Test Files** | 16 Python test files | 252 `test_*.ts` files (252 auto-discovered via `globSync`) |
 
 ### Summary
 
@@ -70,7 +70,7 @@ DEXBot2 is a ground-up rewrite in TypeScript that prioritizes production correct
 | **Process Manager** | Systemd service (Linux) | `dexbot start` (native monolithic launcher, default; `unlock` runtime) — PM2 optional |
 | **External APIs** | CoinGecko, CCXT, Waves | On-chain/pool/Kibana candle inputs; optional CEX synthetic seed generator (`fetch_cex_synthetic_data.ts`) for adapter bootstrap |
 | **Container** | Docker (Ubuntu 18.04) | Docker (multi-stage) |
-| **Testing** | pytest + Docker testnet | Native Node assert (249 `test_*.ts` files; auto-discovered via `globSync`) |
+| **Testing** | pytest + Docker testnet | Native Node assert (252 `test_*.ts` files; auto-discovered via `globSync`) |
 | **CI/CD** | Travis CI, AppVeyor | GitHub Actions / local deterministic script suite |
 | **Packaging** | PyInstaller (Win/Mac/Linux binaries) | npm / `unlock` + optional PM2 ecosystem |
 
@@ -616,7 +616,7 @@ Where:
 ### DEXBot2
 
 - **Framework:** Native Node `assert` module (no external test framework)
-- **249 `test_*.ts` files** auto-discovered via `globSync` (`tests/test_*.ts` + `claw/tests/test_*.ts`), covering:
+- **252 `test_*.ts` files** auto-discovered via `globSync` (`tests/test_*.ts` + `claw/tests/test_*.ts`), covering:
   - Unit tests: accounting, strategy, grid, manager logic
   - Copy-on-Write semantics: COW commits, guards, concurrent fills
   - Edge cases: authoritative full-fill resolution, partial fills, BTS fee accounting, precision
@@ -634,7 +634,7 @@ Where:
 | Feature | DEXBot | DEXBot2 |
 |---|---|---|
 | **Framework** | pytest | Native Node assert |
-| **Test Count** | 16 Python test files | 249 `test_*.ts` files; auto-discovered via `globSync` |
+| **Test Count** | 16 Python test files | 252 `test_*.ts` files; auto-discovered via `globSync` |
 | **Test Types** | Unit + integration | Unit + integration + edge-case + runtime regression |
 | **Testnet Integration** | Yes (Docker) | No (mocks) |
 | **External Dependency** | pytest, Docker | None |
@@ -780,13 +780,13 @@ Where:
 
 | Metric | DEXBot | DEXBot2 |
 |---|---|---|
-| **Release Track** | 1.0.0 | v1.4.19 |
+| **Release Track** | 1.0.0 | v1.4.21 |
 | **Active Since** | ~2018 | December 2025 |
 | **Last Commit** | May 23, 2020 | 2026-08-19 |
 | **Total Commits** | 2281 | 2,034 (v1.4.17) |
 | **Lines of Code** | ~10,846 Python LOC in `dexbot/` | Large TypeScript runtime + adapter + Claw + analysis + tests |
 | **Source Files** | 72 Python files in `dexbot/` | ~527 TS files across the repo |
-| **Test Files** | 16 Python test files | 249 `test_*.ts` files (249 auto-discovered via `globSync`) |
+| **Test Files** | 16 Python test files | 252 `test_*.ts` files (252 auto-discovered via `globSync`) |
 | **Documentation** | Sphinx docs + README | 50+ Markdown docs plus Claw skills/references |
 | **Strategies** | 3 + plugins | 1 |
 | **Max Concurrent Bots** | Many (one process) | Many (one monolithic process by default; per-bot via `--isolated`/PM2) |
@@ -814,7 +814,7 @@ Each bottleneck in DEXBot compounds because they run *serially in sequence* —
 | # | Bottleneck | DEXBot | DEXBot2 (with references) | Multiplier |
 |---|------------|--------|---------------------------|------------|
 | 1 | **RPC queries** | Per-order `get_objects` loop called twice per cycle → 2×N sequential RPCs. | Single batch `get_objects([…])` in `fetchRefBlock` — **`modules/bitshares-native/tx/builder.ts`**; parallel account refresh via `Promise.all` — **`modules/bitshares-native/subscriptions.ts`** (`refreshTasks`, `scanTasks`). | **~400×** |
-| 2 | **Order counting** | Geometric while-loop iterating `price /= 1+increment` ~920 times per call. | O(1) `Math.log` spread-step formula in `calculateSpreadSteps` — **`modules/order/utils/math.ts`**; O(1) spread check in `shouldFlagOutOfSpread` — **`modules/order/utils/order.ts`**. | **~920× CPU** |
+| 2 | **Order counting** | Geometric while-loop iterating `price /= 1+increment` ~920 times per call. | O(1) `Math.log` spread-step formula in `calculateGapSlots` — **`modules/order/utils/math.ts`**; O(1) spread check in `shouldFlagOutOfSpread` — **`modules/order/utils/order.ts`**. | **~920× CPU** |
 | 3 | **Market price** | Fresh `ticker()` RPC inside every order placement. | Cached `centerPrice` served from `botState`, zero RPC per placement — **`market_adapter/market_adapter.ts`** + **`market_adapter/core/market_adapter_service.ts`**. | **~∞ (eliminated)** |
 | 4 | **Account refresh** | Full `_account.refresh()` fetches all orders + balances + history every cycle. | Targeted `set_subscribe_callback` pushing only `OP_FILL_ORDER` ops — **`modules/bitshares-native/subscriptions.ts`** (`refreshSubscriptions`); filtered per-account; no full re-read — **`modules/chain_orders.ts`**. | **~50×** |
 | 5 | **Thread blocking** | `time.sleep(2–6)` on retry blocks the GIL thread entirely. | Async `await sleep()` + `AsyncLock` queue — **`modules/order/async_lock.ts`**; lock guards in **`modules/order/manager.ts`**; backoff in **`modules/order/utils/system.ts`**. | **~100× I/O utilization** |
@@ -838,7 +838,7 @@ DEXBot2 parallel path:  [batch RPC] ─┐
                          [JSON write] ┘
 ```
 
-The Python runtime overhead (≈2× slower than V8 on equivalent CPU work) is the *least* impactful factor here — but it still compounds with everything else. The geometric while-loops and `_calc_increase` iterations all run at Python bytecode speed, thousands of iterations per cycle. DEXBot2 eliminates the iterations entirely with O(1) formulas (`calculateSpreadSteps` in `math.ts`, `shouldFlagOutOfSpread` in `order.ts`) — so the 2× language factor is just insurance on top of the architectural gains.
+The Python runtime overhead (≈2× slower than V8 on equivalent CPU work) is the *least* impactful factor here — but it still compounds with everything else. The geometric while-loops and `_calc_increase` iterations all run at Python bytecode speed, thousands of iterations per cycle. DEXBot2 eliminates the iterations entirely with O(1) formulas (`calculateGapSlots` in `math.ts`, `shouldFlagOutOfSpread` in `order.ts`) — so the 2× language factor is just insurance on top of the architectural gains.
 
 The 500× figure is not theoretical: it materializes in production when higher order counts, slower public nodes, transient block-expiration retries, and wide geometric ranges all hit at once — a scenario DEXBot handles by piling seconds onto seconds, while DEXBot2 absorbs each factor with negligible marginal cost.
 
@@ -886,7 +886,7 @@ The 500× figure is not theoretical: it materializes in production when higher o
 | **Security** | ★★★☆☆ | ★★★★★ (AES-256-GCM, credential daemon, authority resolution) | DEXBot2 |
 | **Ease of Setup** | ★★☆☆☆ (PyQt5/PyInstaller/Systemd dependency hell) | ★★★★★ (`npm i -g dexbot`, zero deps) | DEXBot2 |
 | **Accessibility** | ★★★★★ (GUI) | ★★☆☆☆ (CLI only) | DEXBot |
-| **Testing Depth** | ★★★☆☆ | ★★★★★ (249 test files; focused regressions) | DEXBot2 |
+| **Testing Depth** | ★★★☆☆ | ★★★★★ (252 test files; focused regressions) | DEXBot2 |
 | **Documentation** | ★★★☆☆ | ★★★★★ (architecture/accounting/security/adapter docs) | DEXBot2 |
 | **Dependency Footprint** | ★★☆☆☆ (heavy) | ★★★★★ (0 runtime deps) | DEXBot2 |
 | **Extensibility** | ★★★★★ (plugins) | ★☆☆☆☆ | DEXBot |

package/docs/EVOLUTION.md

@@ -2,14 +2,14 @@
 
 ## Executive Summary
 
-DEXBot2 is a sophisticated decentralized exchange trading bot for the BitShares blockchain. This report documents the complete evolution of the project from its inception in December 2025 through the current 1.4.19 stable release.
+DEXBot2 is a sophisticated decentralized exchange trading bot for the BitShares blockchain. This report documents the complete evolution of the project from its inception in December 2025 through the current 1.4.21 stable release.
 
 ### Key Milestones
 - **Project Inception**: December 2, 2025
-- **Growth Phase**: 2,024 commits over ~8 active months
+- **Growth Phase**: 2,068 commits over ~9 active months
 - **Code Maturity**: Evolution from basic utilities to a ~70,000+ LoC intelligent TypeScript system
-- **Stability**: Progression from manual testing to a suite of 249 automated test files
-- **Releases**: 96 release entries (v0.1.0 to v1.4.19)
+- **Stability**: Progression from manual testing to a suite of 252 automated test files
+- **Releases**: 98 release entries (v0.1.0 to v1.4.21)
 
 ---
 
@@ -61,13 +61,13 @@ v1.4.12 completed the module transition to native ES modules (root + claw `"type
 
 ### Phase 9: Post-ESM Cleanup, Consolidation & Hardening (August 2026)
 
-v1.4.16 centralized all user/runtime state onto a single resolver-derived profiles dir (`~/.config/dexbot2/profiles`) so it survives re-clones and npm updates and never lands in a read-only package dir, turned divergence surplus-cancel + hole-create pairs into in-place order rotations, and added an npm auto-update flow. v1.4.17 consolidated duplicate code (EC math, Base58Check, settings merge, asset resolution), trimmed `modules/types.ts` from 875 lines to the Order union, purged dead exports, and centralized the analysis tooling under strict TypeScript. v1.4.19 capped COW broadcasts at `MAX_OPS_PER_BROADCAST` (4) with chunked retry-on-uncertain broadcasting, fixed a spread-collapse regression via the shared `isSlotInRail` helper, lowered the AMA slope grid-reset threshold to 8, and added editor price feedback plus a `docs/LIFECYCLE.md` onboarding walkthrough.
+v1.4.16 centralized all user/runtime state onto a single resolver-derived profiles dir (`~/.config/dexbot2/profiles`) so it survives re-clones and npm updates and never lands in a read-only package dir, turned divergence surplus-cancel + hole-create pairs into in-place order rotations, and added an npm auto-update flow. v1.4.17 consolidated duplicate code (EC math, Base58Check, settings merge, asset resolution), trimmed `modules/types.ts` from 875 lines to the Order union, purged dead exports, and centralized the analysis tooling under strict TypeScript. v1.4.19 capped COW broadcasts at `MAX_OPS_PER_BROADCAST` (4) with chunked retry-on-uncertain broadcasting, fixed a spread-collapse regression via the shared `isSlotInRail` helper, lowered the AMA slope grid-reset threshold to 8, and added editor price feedback plus a `docs/LIFECYCLE.md` onboarding walkthrough. v1.4.20 hardened spread-correction boundary promotion with a MIN_SPREAD_ORDERS reserve floor and commit-time validation, gated persisted-boundary restore against self-legalizing overrun poison, aligned dynamic-weight clip parity between the live service and the research chart, centralized chart slider ranges, and moved all analysis outputs onto the central path resolver. v1.4.21 fixed silent-failure runtime defects surfaced by a modules-wide audit (NaN fund-invariant tolerance, always-flush fill store, double-decremented fill guard, config num() defaults), aligned the two boundary writers on a shared sell-rail ceiling enforced at commit and restore time, deduplicated claw shared logic while hardening error paths (secret redaction, EPIPE guards, spawn-error surfacing), corrected launcher supervisor/runtime lifecycle bugs, made the browser storage adapter persist deletions and flush debounced, extended editor green/red live color feedback to funds and price inputs, stopped docs/error messages from hardcoding repo-relative profile paths, and pruned dead code across analysis/market_adapter/claw/logging.
 
 ---
 
 ## Development Statistics
 
-The project has accumulated 249 automated test files across 95 release entries. See the **Version History** below for a per-release commit breakdown.
+The project has accumulated 252 automated test files across 98 release entries. See the **Version History** below for a per-release commit breakdown.
 
 ---
 
@@ -86,7 +86,7 @@ The project has accumulated 249 automated test files across 95 release entries.
 
 ## Documentation & Testing
 
-Evolved from a basic README to a comprehensive framework (50+ docs entries, 80%+ JSDoc coverage, AGENTS.md). Testing matured from manual blockchain trials → Jest → lightweight Node.js assert across a 249-file suite covering unit, integration, simulation, and COW architectural guard tests.
+Evolved from a basic README to a comprehensive framework (50+ docs entries, 80%+ JSDoc coverage, AGENTS.md). Testing matured from manual blockchain trials → Jest → lightweight Node.js assert across a 252-file suite covering unit, integration, simulation, and COW architectural guard tests.
 
 ---
 
@@ -131,11 +131,13 @@ Compact, era-level view; per-release commit detail lives in [CHANGELOG.md](../CH
 | v1.3.3 → v1.4.8 | 74 | CJS→ESM completion, concurrency correction, uncertain-broadcast safety, truncated-read ambiguity |
 | v1.4.8 → v1.4.13 | 45 | Native ESM runtime, broadcast serialization, onboarding |
 | v1.4.13 → v1.4.19 | 36 | Profile-state centralization, code consolidation, per-broadcast op cap |
+| v1.4.19 → v1.4.20 | 5 | Grid boundary promotion hardening, recovery poison gate, analysis output centralization |
+| v1.4.20 → v1.4.21 | 15 | Runtime audit fixes, claw dedup hardening, boundary ceiling alignment, editor color feedback |
 
 ---
 
 **Report Originally Generated**: February 19, 2026
-**Last Updated**: August 21, 2026 (v1.4.19)
-**Total Commits**: 2,044
-**Date Range**: December 2, 2025 – August 21, 2026
+**Last Updated**: August 24, 2026 (v1.4.21)
+**Total Commits**: 2,068
+**Date Range**: December 2, 2025 – August 24, 2026
 **Repository**: DEXBot2 (BitShares DEX Trading Bot)

package/docs/FUND_MOVEMENT_AND_ACCOUNTING.md

@@ -52,15 +52,13 @@ funds.allocated → _getSizingContext()       (budget for spread correction)
 **Key points:**
 - `botFunds` percentage applies to **total** capital (free + locked in orders), not just free. A bot at `"50%"` gets half of everything, not half of what's currently idle.
 - `funds.allocated` is the ceiling for each side. Existing orders already consume part of it; the remaining free portion is available for new placements.
-- The downstream `applyBotFundsAllocation()` (`manager.ts:953`) also caps `funds.available` to `<= allocated` as a safety net, but the primary budget chokepoint is `getSideBudget` / `_getSizingContext` reading `allocated` directly (v1.2.6).
+- The downstream `applyBotFundsAllocation()` (`manager.ts:949`) also caps `funds.available` to `<= allocated` as a safety net, but the primary budget chokepoint is `getSideBudget` / `_getSizingContext` reading `allocated` directly (v1.2.6).
 
 ---
 
 ### 1.4 Mixed Order Fund Validation
 
-**Problem Fixed**: Early batch builders ran a single fund check keyed off the first order's side, so a mixed BUY/SELL batch could trip a false "insufficient funds" warning even when each side had ample capital — the BUY check was applied to SELL orders (or vice versa).
-
-**Solution**: Per-asset validation using a signed-delta water-mark. The current validator tracks the **peak** running requirement per asset (not a side lump sum), so BUY and SELL ops in the same batch are validated independently against their own free balance.
+Mixed BUY/SELL batches are validated per asset using a signed-delta **peak** running requirement (not a side lump sum), so BUY and SELL ops in the same batch are checked independently against their own free balance.
 
 #### Fund Availability Checks by Asset
 
@@ -69,7 +67,7 @@ funds.allocated → _getSizingContext()       (budget for spread correction)
 
 #### Implementation Location
 
-File: `modules/dexbot_cow_runtime.ts` — `validateOperationFunds()` (line 1309), called from the COW batch broadcast path at line 2648. `modules/dexbot_class.ts` exposes a thin wrapper `_validateOperationFunds()` (line 1165).
+File: `modules/dexbot_cow_runtime.ts` — `validateOperationFunds()` (line 1488), called from the COW batch broadcast path at line 2908. `modules/dexbot_class.ts` exposes a thin wrapper `_validateOperationFunds()` (line 1122).
 
 ```javascript
 // Per-asset peak requirement vs. quantized chain-free snapshot.
@@ -112,19 +110,9 @@ See [developer_guide.md#order-state-helper-functions](developer_guide.md#order-s
 
 ### 1.5 Fill Batch Processing & Timeline
 
-#### Problem Solved
-
-Previously, fills were processed one-at-a-time (~3s per broadcast). A burst of 29 fills in the Feb 7 market crash took ~90 seconds, during which:
-- Market prices moved significantly
-- Orders became stale (filled on-chain but not yet synced)
-- Orphan fills were created (fill events for orders no longer on-chain)
-- Fund tracking diverged from blockchain reality
-
-**Impact**: The extended 90s window meant the bot couldn't react to market moves, creating a cascading failure.
-
-#### Solution: Fixed-Cap Batch Fill Processing
+#### Fixed-Cap Batch Fill Processing
 
-**Mechanism**: Fill events arrive via `modules/dexbot_fill_runtime.ts` (the fill-runtime module), which pushes them into `bot._incomingFillQueue` (declared in `modules/dexbot_class.ts`). The drain loop in `dexbot_fill_runtime.ts` then chunks the queue into capped batches and calls `modules/order/manager.ts::processFilledOrders` (line 1442) once per chunk to run the full rebalance pipeline.
+**Mechanism**: Fill events arrive via `modules/dexbot_fill_runtime.ts` (the fill-runtime module), which pushes them into `bot._incomingFillQueue` (declared in `modules/dexbot_class.ts`). The drain loop in `dexbot_fill_runtime.ts` then chunks the queue into capped batches and calls `modules/order/manager.ts::processFilledOrders` (line 1438) once per chunk to run the full rebalance pipeline.
 
 **Batch Sizing Algorithm**: A single cap-based batch size (`FILL_PROCESSING.MAX_FILL_BATCH_SIZE`): a queue depth of 4 or fewer is processed as one unified batch; deeper queues are chunked into repeated batches of 4 (the last chunk may be smaller).
 
@@ -176,9 +164,7 @@ IF ratio >= GRID_REGENERATION_PERCENTAGE (default: 3%):
 
 #### Recovery Retry System
 
-**Problem**: One-shot `_recoveryAttempted` boolean flag meant permanent lockup if recovery failed once.
-
-**New Behavior**: Count+time-based retry system with periodic reset.
+Recovery uses a count+time-based retry system with periodic reset, so a single failed recovery attempt never locks out future retries.
 
 **State Machine**:
 ```
@@ -221,25 +207,14 @@ When a batch fails because an on-chain order no longer exists, the cleanup relea
 
 ### 1.6 Remainder Accuracy During Capped Resize
 
-#### Problem Fixed
-
-When grid resize was capped by available funds, the accounting system needed to track what portion of the ideal grid went unallocated. This required careful per-slot tracking to distinguish between:
-- **Fully allocated slots**: received their ideal size (no remainder)
-- **Fund-capped slots**: received less than ideal because available funds ran out mid-allocation
+When grid resize is capped by available funds, accounting tracks what portion of the ideal grid went unallocated via per-slot tracking:
+- **Fully allocated slots**: receive their ideal size (no remainder)
+- **Fund-capped slots**: receive less than ideal because available funds ran out mid-allocation
 
-Without per-slot tracking, the remainder was computed from totals, which overstated it when some slots were fully allocated and others were capped.
+Computing the remainder from totals instead overstates it when some slots are fully allocated and others are capped.
 
-#### Solution: Per-Slot Tracking
-
-**Old Behavior** (Incorrect):
-```javascript
-// Compute unallocated remainder from ideal sizes
-const remainder = totalIdealSizes - totalAllocatedSizes;
-// Problem: If actual allocation capped at 80% due to insufficient funds,
-// this uses 100% ideal in calculation → remainder overstated
-```
+#### Per-Slot Tracking
 
-**New Behavior** (Correct):
 ```javascript
 // Track per-slot applied sizes
 const appliedSizes = [];
@@ -415,25 +390,9 @@ targetSlot.orderId = newOrderId;
 
 ### 3.6 Orphan-Fill Deduplication & Double-Credit Prevention
 
-**Location**: `modules/dexbot_class.ts` — constructor, `_recoverExplicitStaleOrders()` (line 445), orphan-fill guard in the fill drain loop (line ~1633), and pruning pass after each cycle (line ~1919).
-
-#### Problem Solved
-
-During Feb 7 market crash, stale-order batch failures cascaded into double-crediting:
-
-**Scenario**:
-1. Batch operation scheduled with 12 orders
-2. Order X is on-chain, included in batch
-3. Between sync and broadcast, order X fills on market (stale order)
-4. Batch execution fails: "Limit order X does not exist"
-5. Error handler: Clean up grid slot X, release funds to `chainFree`
-6. Meanwhile, fill event arrives: "Order X was filled at price Y for amount Z"
-7. Orphan-fill handler: Credits proceeds to `chainFree` AGAIN
-8. **Result**: Double-credit of proceeds, inflated `chainTotal`, fund drift
+**Location**: `modules/dexbot_class.ts` — constructor, `_recoverExplicitStaleOrders()` (line 448); orphan-fill guard in the fill drain loop and pruning pass after each cycle now live in `modules/dexbot_fill_runtime.ts` (guard ~lines 628-646, pruning ~lines 934-946).
 
-**In Crash Numbers**: 7 orphan fills × ~700 BTS = ~4,600 BTS inflated → cascaded to 47,842 BTS total drift.
-
-#### Solution: Stale-Cleaned Order ID Tracking with TTL + Recycled-Slot Guard
+#### Stale-Cleaned Order ID Tracking with TTL + Recycled-Slot Guard
 
 **Mechanism**: Track which orders were cleaned up during batch failure recovery using timestamp + grid-slot retention.
 
@@ -566,7 +525,7 @@ BitShares charges fees for `limit_order_create` and `limit_order_cancel`.
     2.  If sufficient `chainFree` available: deduct full amount atomically.
     3.  If insufficient: defer settlement and retry when funds become available.
 
--   **Adoption fee parity (1.4.8):** COW chain-adoption paths charge fees exactly like the normal open-orders loop. `adoptPlacedBatchFromChain` (refused-commit and poll-confirmed paths) and the startup uncertain-create adoption apply the create/cancel/update fees via `_applySync` — previously create-only charging let optimistic BTS drift when a batch's orders were adopted without fee accounting.
+-   **Adoption fee parity:** COW chain-adoption paths charge fees exactly like the normal open-orders loop. `adoptPlacedBatchFromChain` (refused-commit and poll-confirmed paths) and the startup uncertain-create adoption apply the create/cancel/update fees via `_applySync`.
 
 -   **Safe fee lookup:** `processBatchResults` uses `getAssetFeesSafe('BTS')` with zero-fee fallbacks — the throwing variant can no longer hard-fail a whole batch after a successful commit (`modules/dexbot_cow_runtime.ts`).
 
@@ -588,7 +547,7 @@ These are deducted from the *proceeds* of a fill.
 
 For BTS fees, the system returns a structured object (not a simple number) with multiple fields for accounting precision.
 
-**Location**: `modules/order/utils/math.ts::getAssetFees()` (line 312). The fee cache itself is populated by `modules/order/utils/system.ts::initializeFeeCache()` (line 631).
+**Location**: `modules/order/utils/math.ts::getAssetFees()` (line 312). The fee cache itself is populated by `modules/order/utils/system.ts::initializeFeeCache()` (line 665).
 
 #### BTS Fee Object (Always Object)
 
@@ -651,10 +610,6 @@ const createFee = feeInfo.createFee;   // BTS only
 
 ### 5.4 BUY Side Sizing & Fee Accounting
 
-**Problem Fixed**: BUY side fee calculations incorrectly applied fees to base asset instead of quote asset.
-
-**Solution**: Corrected fee accounting with proper asset assignment.
-
 #### Fee Application by Side
 
 | Side | Asset | Calculation | Notes |
@@ -702,13 +657,11 @@ For BUY orders that are makers:
 
 ### 5.5 Precision & Quantization
 
-**Problem**: Floating-point arithmetic accumulates rounding errors over many calculations. After dozens of order size calculations, price derivations, and fund allocations, float values drift from their true blockchain integer representations, causing mismatches between internal state and on-chain reality.
-
-**Solution**: Centralized quantization utilities that eliminate float accumulation by round-tripping through blockchain integer representation.
+Floating-point arithmetic drifts from true blockchain integer representations over many order-size calculations, price derivations, and fund allocations. Quantization eliminates this accumulation by round-tripping every value through its blockchain integer form.
 
 #### 5.5.1 Core Quantization Functions
 
-**Location**: `modules/order/utils/math.ts` (line 260)
+**Location**: `modules/order/utils/math.ts` (line 265)
 
 ##### `quantizeFloat(value, precision)` - Eliminate Accumulation Errors
 
@@ -771,21 +724,13 @@ const normalized = normalizeInt(currentSizeInt, 8);
 - Normalizing fund totals before invariant checks
 - Preparing sizes for blockchain transaction encoding
 
-#### 5.5.2 Consolidation Impact
+#### 5.5.2 Consolidated Quantization
 
-Previously, five separate quantization implementations existed:
-- `dexbot_class.ts` - Manual rounding logic
-- `order.ts` - Custom precision handling
-- `strategy.ts` - Divergent rounding approach
-- `chain_orders.ts` - Different quantization pattern
-- `export.ts` - Isolated float conversions
+Quantization has a single source of truth: `quantizeFloat()` in `modules/order/utils/math.ts`.
 
-**After Consolidation:**
-✅ Single source of truth (`math.ts`)
 ✅ Consistent precision handling across all modules
 ✅ Reduced regression risk (tested once, used everywhere)
-✅ Eliminated subtle float accumulation bugs
-✅ All 34+ test suites pass with zero regressions
+✅ No subtle float accumulation bugs from divergent rounding paths
 
 #### 5.5.3 Precision Best Practices
 
@@ -818,7 +763,7 @@ This prevents the bug where `available = chainFree + required` created a tautolo
 
 ## 6. Safety & Invariants
 
-The `Accountant` enforces strict mathematical invariants to detect bugs or manual interference. Invariants are checked by `_verifyFundInvariants()` (`modules/order/accounting.ts` line 502) after every blockchain sync cycle. The verification reads from a snapshot captured under `_fundLock` — `actualBuy`/`actualSell` are captured at snapshot time, not read live outside the lock, closing a TOCTOU window. When a violation is detected, the system logs a `CRITICAL` error and attempts automatic recovery via `manager.accountant.recalculateFunds()` (`modules/order/accounting.ts` line 347, delegated from `modules/order/manager.ts` lines 981–990) — resetting internal state to match on-chain reality. If the grid lock is held (mid-rebalance), recovery is deferred until the lock is released. The bot continues operating throughout; it does **not** halt on invariant violations.
+The `Accountant` enforces strict mathematical invariants to detect bugs or manual interference. Invariants are checked by `_verifyFundInvariants()` (`modules/order/accounting.ts` line 501) after every blockchain sync cycle. The verification reads from a snapshot captured under `_fundLock` — `actualBuy`/`actualSell` are captured at snapshot time, not read live outside the lock, closing a TOCTOU window. When a violation is detected, the system logs a `CRITICAL` error and attempts automatic recovery via `manager.accountant.recalculateFunds()` (`modules/order/accounting.ts` line 346, delegated from `modules/order/manager.ts` lines 981–990) — resetting internal state to match on-chain reality. If the grid lock is held (mid-rebalance), recovery is deferred until the lock is released. The bot continues operating throughout; it does **not** halt on invariant violations.
 
 ### 6.1 The Equality Invariant
 Total funds on chain must equal free plus committed.
@@ -834,18 +779,17 @@ A violation here means the grid has allocated more capital than actually exists
 
 ### 6.3 Race Condition Protection (TOCTOU)
 To prevent "Time-of-Check to Time-of-Use" errors:
-1.  **Locking:** `AsyncLock` (re-entrant) prevents concurrent updates to the same order. Nested `acquire()` from the same execution context runs the callback directly instead of queueing, eliminating the `fillLockAlreadyHeld` parameter that previously threaded through 25+ call sites.
+1.  **Locking:** `AsyncLock` (re-entrant) prevents concurrent updates to the same order. Nested `acquire()` from the same execution context runs the callback directly instead of queueing.
 2.  **Atomic Deduct:** `tryDeductFromChainFree` checks *and* subtracts in a single synchronous step.
 3.  **Bootstrapping:** Fills arriving during startup (`isBootstrapping=true`) are queued until the grid is fully reconciled ([GRID_RECONCILE.md](GRID_RECONCILE.md)).
 
-### 6.4 Stale Accounting & Fee Over-Credit Guards (v1.2.1)
-Two additional accounting hardening measures added in v1.2.1:
+### 6.4 Stale Accounting & Fee Over-Credit Guards
 
-**Stale `accountTotals` no longer HARD-ABORTs COW commit.** Transient staleness (e.g., the periodic balance fetch overlaps with a COW commit) logs a `WARN` and schedules recovery instead of throwing `ACCOUNTING_COMMITMENT_FAILED`. Totals are also refreshed after bootstrap to prevent a spurious full recovery on the first maintenance cycle.
+**Stale `accountTotals` does not abort COW commit.** Transient staleness (e.g., the periodic balance fetch overlaps with a COW commit) logs a `WARN` and schedules recovery instead of throwing `ACCOUNTING_COMMITMENT_FAILED`. Totals are also refreshed after bootstrap to prevent a spurious full recovery on the first maintenance cycle.
 
-**Fee-deduction failure escalates to `error`.** When `getAssetFees` throws during fill processing (e.g., network blip), `_deductFeesFromProceeds` previously returned raw proceeds without deduction and logged at `warn` level — silently inflating `accountTotals` over time. The log is now `error` with explicit "fund tracking will over-credit" language so operators can detect the drift source in production logs.
+**Fee-deduction failure logs at `error`.** When `getAssetFees` throws during fill processing (e.g., network blip), `_deductFeesFromProceeds` skips the deduction and logs at `error` with explicit "fund tracking will over-credit" language so operators can detect the drift source in production logs.
 
-**TOCTOU in `processFillAccounting`.** `_buildBtsDeferredRefundAdjustment` reads `btsFeeState` from `mgr.orders`, but the order lock was acquired after accounting ran. Fixed by acquiring the lock first, then running `processFillAccounting` under the lock. This fix was also ported to POST-RESET and BOOTSTRAP tracked-fill accounting paths.
+**TOCTOU protection in `processFillAccounting`.** `_buildBtsDeferredRefundAdjustment` reads `btsFeeState` from `mgr.orders` while the order lock is held — the lock is acquired before accounting runs, and the POST-RESET and BOOTSTRAP tracked-fill accounting paths follow the same locking pattern.
 
 ---
-*Technical Reference for DEXBot2 v1.4.19 release*
+*Technical Reference for DEXBot2 v1.4.21 release*

package/docs/GRID_RECALCULATION.md

@@ -397,7 +397,7 @@ and dedupes while one resync is already pending or running.
 - Fill/rotation operations are causing noticeable size errors
 
 **Disable (set to 0) if:**
-- You want to rely ONLY on AMA triggers (Issue #5: RMS Divergence Check Disabling)
+- You want to rely ONLY on AMA triggers
 - You want to prevent automatic full resync from divergence alone
 - You manually trigger full grid resyncs through other mechanisms
 
@@ -596,18 +596,6 @@ Removed trigger file.
 
 ---
 
-## Related Issues
-
-> Issue numbers below refer to the project's historical internal tracker and
-> are kept here for archival context. They are not always resolvable against
-> external issue trackers.
-
-- **Issue #5:** RMS Divergence Check Disabling — Ability to set `RMS_PERCENTAGE: 0` to disable checks
-- **Feature:** AMA Integration — AMA-derived center snapshots are already used for market-adapter-triggered grid recentering
-- **Issue #1:** Fund Validation Bug — Fixed validation logic for order batch placement
-
----
-
 ## References
 
 - `modules/constants.ts` — Default configuration values