toga-ai 1.0.511 → 1.0.512

package/knowledge/2.0/apps/_underscore/features/error-reporting-issue-event.md

@@ -364,6 +364,18 @@ in `_underscore` and `library` in the **same** release, and the `/errors` consol
   `dbchanges2/Logs_Client/2026-08-01a - Drop Error table.sql`. Do not reintroduce it.
 - **Shared-schema migration collision.** Because these tables live in the *shared* Core Logs
   DB, only **one** migration may `CREATE` them; later migrations only `ALTER`.
+- **The `dbchanges2/Logs/` migration files LAG the deployed schema — read the live cluster, not the
+  migrations.** `dbchanges2/Logs/2026-07-30a - Error reporting Issues and Events.sql` does **not**
+  describe the live `Logs.Issue`: notably the identity column is **`issueKey` (varchar 128)**, not
+  `code`. Live columns include `reference`, `dtLastOccurred`, `isManaged`, `urgency`,
+  `minimumUrgency`, `consecutiveQuietWindows`, `subject`, `errorMessage`, `occurrences`,
+  `totalOccurrences`, `recurrenceCount`, `dtLastNotified`, `notificationCount`, `dtAcknowledged`,
+  `dtMutedUntil`, `createsClickupTask`, `clickupPriority`, `clickupAssigneeId`. Query
+  `information_schema` (or `SHOW TABLES` / `SHOW COLUMNS`) against the live Logs cluster **before**
+  writing any SQL against this feature — reading the migrations instead produced confidently wrong
+  conclusions twice on 2026-08-04. Same trap for the **singular** table names
+  (`Issue`/`Event`/`IssueFingerprint`/`IssueAreaOwner`/`IssueClickupTask`/`IssueEmailAddress`,
+  renamed by `2026-08-01a` and confirmed applied in prod).
 - **Pre-existing committed secrets in this area (not fixed — out of scope, report only).**
   A ClickUp API token at `_underscore/Component/Api/Clickup/Clickup.php:5`; SES SMTP
   credentials at `_underscore/Email.php:11-12`; a named individual's Sentry personal access
@@ -418,6 +430,12 @@ clientUserId). **Neither was built.** As built instead:
 
 ## Change history
 
+- 2026-08-04 — **Correction/diagnostic:** recorded that the `dbchanges2/Logs/` migration files lag
+  the deployed schema — `2026-07-30a` describes a `code` identity column, but live `Logs.Issue` is
+  keyed on **`issueKey` (varchar 128)** — and listed the live column set. Query
+  `information_schema` against the live Logs cluster before writing SQL here; trusting the
+  migrations (and forgetting the applied singular rename) produced two confidently wrong
+  conclusions in one session. No code change. (dfranks)
 - 2026-08-04 (later, **uncommitted/undeployed** at time of writing) — **Reference encoding reworked
   to base 32** over `'0123456789ACDEFGHJKMNPQRSTUVWXYZ'` (all digits kept; `O`/`I`/`L`/`B` dropped
   as impersonators of `0`/`1`/`1`/`8`), replacing the base-26 decimal-counter+letter scheme, with

package/knowledge/2.0/apps/worker2/features/error-escalation-cron.md

@@ -7,7 +7,7 @@ client: shared
 type: feature
 status: active
 updated: 2026-08-04
-owners: ["jcardinal"]
+owners: ["jcardinal", "dfranks"]
 files:
   - worker2/Worker/Infrastructure/Errors.php
   - worker2/Worker/Notification/Email.php
@@ -28,6 +28,7 @@ related:
   - ./clickup-project-routing.md
   - ./notification-email.md
   - ./notification-email-template.md
+  - ../../../clients/compass-usa/features/mits-sales-order-transmission-alerting.md
 ---
 
 ## Summary
@@ -263,6 +264,22 @@ this controller needs the same treatment.**
   issue gets a **neutral** subject rather than another client's raw error text.
 - **`IssueEmailAddress.clientId = 0` means "all clients"** and requires an explicit
   confirmation flag in the Tools console; list views render it as **"ALL CLIENTS"**.
+- **A recipient scoped to a specific `clientId` delivers NOTHING for occurrences recorded without
+  client attribution.** `loadRecentClientIds()` maps an occurrence whose `clientId` is NULL to
+  **`0`**, and `resolveRecipientsForClient()` resolves scope `0` **plus** the erroring client. So
+  the rule is: scope to **`0`** when the `issueKey` is raised from a **single client-specific code
+  path** (a Compass-only cron, say); **never** scope to `0` for an `issueKey` raised from shared
+  code, which would leak one client's alerts to another's recipients.
+- **Configuring a recipient is never retroactive.** `shouldNotify()` notifies on first occurrence
+  (`notificationCount` 0), on escalation, or when the backed-off reminder interval has elapsed —
+  and returns **false when the current window has zero events**. The first email therefore lands on
+  the **next live occurrence**, not on the backlog. Do not read silence right after a config change
+  as a failure.
+- **Adding a business recipient is purely additive — it cannot switch ClickUp ticketing off.**
+  Routing is **both**, not either: `if ($recipients !== null)` emails the business recipients and
+  `if ($issue->createsClickupTask)` tickets the dev team, evaluated **independently**. The in-code
+  comment marking this records the earlier `elseif` form, where adding one recipient silently
+  disabled an issue's dev-team ticket.
 - **The ClickUp webhook has no HMAC signature verification** (pre-existing, unfixed). A forged
   POST can mark issues acknowledged — freezing the neglect axis — or resolve episodes. That is
   alert suppression against the error-reporting system itself. Fix before relying on this
@@ -273,6 +290,15 @@ this controller needs the same treatment.**
 
 ## Change history
 
+- 2026-08-04 (verified against `origin/_production` + the prod Logs cluster) — Recorded the
+  **recipient-routing semantics** a developer needs before configuring a business alert: routing is
+  **both** business email and dev-team ClickUp, evaluated independently (adding a recipient is
+  purely additive); `IssueEmailAddress.clientId` scoping means a client-scoped row delivers nothing
+  for occurrences with NULL `clientId` (mapped to `0`), so single-client `issueKey`s scope to `0`
+  and shared-code `issueKey`s never do; and `shouldNotify()` makes configuration
+  **non-retroactive** — the first email lands on the next live occurrence. Also noted that prod
+  `Logs.IssueEmailAddress` was still empty, making Compass's MITS rejection the platform's first
+  business-routed issue. No code change. (dfranks)
 - 2026-08-04 (later, **uncommitted/undeployed** at time of writing) — **Fixed: escalation email had
   never been sent, ever**, for two independent reasons. (A) `processTechnicalIssue()` returned early
   when the ClickUp task would read identically, and that return sat *above* the `if ($shouldNotify)`

package/knowledge/INDEX.md

@@ -39,7 +39,7 @@ _Auto-generated by `knowledge.js index`. Do not hand-edit._
 - **togatech** (TOGA Technology Website) — 4 doc(s) → [standalone/apps/togatech/INDEX.md](standalone/apps/togatech/INDEX.md)
 - **websocket** (WebSocket Server) — 2 doc(s) → [standalone/apps/websocket/INDEX.md](standalone/apps/websocket/INDEX.md)
 - **forward** (Forwarder) — 3 doc(s) → [standalone/apps/forward/INDEX.md](standalone/apps/forward/INDEX.md)
-- **claude** (Claude Harness) — 3 doc(s) → [standalone/apps/claude/INDEX.md](standalone/apps/claude/INDEX.md)
+- **claude** (Claude Harness) — 4 doc(s) → [standalone/apps/claude/INDEX.md](standalone/apps/claude/INDEX.md)
 
 ## Clients
 

package/knowledge/clients/compass-usa/INDEX.md

@@ -8,6 +8,7 @@
 | [Compass: Item-Fulfillment TableViews (for-sales-order-items & for-sales-orders, tracking via bridge)](features/item-fulfillment-tracking-tableview.md) | 2.0 | Two sibling Compass TableViews in `Client_Compass` display fulfilled items in toga2-supply, both driven by `TableViews` / `TableViewJoins` / `TableViewFields` c | dbchanges2/Client_Compass/2026-06-10 - ItemFulfillmentsForSalesOrderItemsTableView.sql, dbchanges2/Client_Compass/2026-06-11 - ItemFulfillmentsForSalesOrdersTableView.sql, dbchanges2/Client_Compass/2026-06-15a - FixItemFulfillmentTrackingNumberJoins.sql, dbchanges2/Client/2026-07-15a - ExcludeFeeItemsFromItemFulfillmentsForSalesOrdersView.sql |
 | [Compass MITS PO → SO Item Linking](features/mits-po-to-so-item-linking.md) | 2.0 | MITS sends Compass inbound Purchase Orders (`POST /v2/purchase-orders`) against a Sales Order (`mitsSalesOrder`). | _underscore/Model/Compass/PurchaseOrder.php, worker/crons/toga2/compass/workflow/3a_import_office_depot_purchase_orders.php |
 | [Compass MITS PO Transmission to Vendors](features/mits-po-transmission-to-vendors.md) | 2.0 | The 1.0 worker cron `2_transmit_mits_purchase_orders_to_vendors.php` transmits Compass PurchaseOrders to their vendors (Office Depot, Strategic Systems, Compass | worker/crons/toga2/compass/workflow/2_transmit_mits_purchase_orders_to_vendors.php, worker/crons/toga2/compasscanada/workflow/2_transmit_mits_purchase_orders_to_vendors.php, library/app/client/compass.php |
+| [Compass MITS Sales-Order Transmission — Rejection Alerting (Issue/Event, not email-in-cron)](features/mits-sales-order-transmission-alerting.md) | 2.0 | When MITS **rejects** a Compass sales order transmitted by the 1.0 cron `1_transmit_compass_sales_orders_to_mits.php`, the alert is no longer an email built ins | worker/crons/toga2/compass/workflow/1_transmit_compass_sales_orders_to_mits.php, worker2/Worker/Infrastructure/Errors.php, _underscore/Model/Core/Logs/Issue.php, _underscore/Model/Core/Logs/IssueEmailAddress.php, dbchanges2/Logs/2026-08-04a - MITS rejection business recipients.sql |
 | [Compass MR/MA Order Auto-Approval & Status Gate](features/mr-ma-order-approval-and-status.md) | 2.0 | Compass **MR** and **MA** sales orders are system-generated from the MITS / Office Depot EDI pipeline (they do not originate as user-entered SA orders) and must | _underscore/Model/Compass/SalesOrder.php, _underscore/Model/Compass/PurchaseOrder.php, worker/crons/toga2/compass/workflow/3a_import_office_depot_purchase_orders.php |
 | [Compass USA](profile.md) | 2.0 | Compass USA is a TOGA client running a multi-tier supply-chain commerce operation. |  |
 | [Compass Cross-Kit Bundle Corruption — Detection & Repair](workflows/cross-kit-bundle-corruption.md) | 2.0 | A frontend regression in `toga2-commerce`'s edit-order bundle submission mis-attributed bundle (kit) line items and **fees/warranties** to the **wrong kit**, pe | src/api/syncSalesOrderItemsFromLocalStorageCartToApi.ts |

package/knowledge/clients/compass-usa/features/mits-sales-order-transmission-alerting.md

@@ -0,0 +1,89 @@
+---
+title: Compass MITS Sales-Order Transmission — Rejection Alerting (Issue/Event, not email-in-cron)
+framework: "2.0"
+project: _Underscore
+client: compass-usa
+type: client-feature
+status: active
+updated: 2026-08-04
+owners: ["dfranks"]
+files:
+  - worker/crons/toga2/compass/workflow/1_transmit_compass_sales_orders_to_mits.php
+  - worker2/Worker/Infrastructure/Errors.php
+  - _underscore/Model/Core/Logs/Issue.php
+  - _underscore/Model/Core/Logs/IssueEmailAddress.php
+  - dbchanges2/Logs/2026-08-04a - MITS rejection business recipients.sql
+related:
+  - mits-po-transmission-to-vendors.md
+  - ../../../2.0/apps/worker2/features/error-escalation-cron.md
+  - ../../../2.0/apps/_underscore/features/error-reporting-issue-event.md
+  - ../../../1.0/apps/library/features/error-capture-1-0.md
+---
+
+## Summary
+
+When MITS **rejects** a Compass sales order transmitted by the 1.0 cron
+`1_transmit_compass_sales_orders_to_mits.php`, the alert is no longer an email built inside the
+cron. It is a **business exception** raised into the shared Logs Issue/Event pipeline, and its
+recipients are **rows in `Logs.IssueEmailAddress`**.
+
+**Consequence for anyone handed a ticket like "add <person> to the MITS error emails": there is
+no PHP to change.** It is a one-row data change delivered as a `dbchanges2/Logs/` migration.
+
+## How it works
+
+- The cron raises
+  `throw new App_Exception_Business('COMPASS_SALES_ORDER_MITS_TRANSMIT_REJECTED', …, URGENCY_HIGH)`
+  on a MITS rejection. (Before 2026-08-03 this block was a hardcoded `new App_Email_Agilant()`
+  with `addTo('devteam@togatech.com')`.)
+- Capture creates/dedupes a `Logs.Issue` keyed on the `issueKey`, and the worker2
+  `Infrastructure/Errors/Escalate` cron does all outbound notification — business email to the
+  configured recipients **and**, independently, the dev-team ClickUp task. See
+  [error-escalation-cron](../../../2.0/apps/worker2/features/error-escalation-cron.md).
+- Recipients live in `Logs.IssueEmailAddress` (FK `issueId` → `Logs.Issue`), scoped by
+  `clientId` (`0` = all clients).
+
+### Adding a recipient (the correct pattern)
+
+Write a `dbchanges2/Logs/YYYY-MM-DD<letter>` migration that:
+
+1. **Looks the issue up by `issueKey`**, never by its `id`. The id is assigned on the issue's
+   **first occurrence** and is therefore not stable across environments.
+2. Uses **`clientId = 0`**. This `issueKey` is raised from a Compass-only code path, and an
+   occurrence recorded without client attribution resolves against scope `0` only — a row scoped
+   to the Compass `clientId` can deliver nothing. (Never use `0` for an `issueKey` raised from
+   shared code — that is a cross-tenant leak.)
+3. Is **idempotent** via `INSERT … WHERE NOT EXISTS`, mirroring the unique key
+   `uq_issue_client_address_type`.
+
+Configuration is **never retroactive**: `shouldNotify()` fires on the first occurrence, on
+escalation, or when the backed-off reminder interval elapses, and returns false for a window with
+zero events. The first email lands on the **next live rejection**, not on existing history.
+
+## Gotchas / known issues
+
+- **Do not "modernize" this alert back into the cron.** Migrating it to
+  `_Model_Client_EmailTemplate::send` (or any in-cron email) reverts the move to the Issue/Event
+  pipeline and removes the recipient configuration surface entirely.
+- **Never re-introduce the payload dump into the alert.** The original email included the
+  transmission payload, which carries ship-to addressee, attention line, street address and phone
+  number. The Issue message is stored in a table **shared across every client**, displayed to
+  internal staff in the Tools `/errors` console, pushed to ClickUp, and emailed — so the payload
+  was deliberately removed. Alerts here name the order, not the person.
+- **The MITS cron holds its API credentials inline in the file.** Treat that file as
+  secret-bearing when copying snippets anywhere (tickets, PR bodies, knowledge docs).
+- **Adding a recipient does not turn ClickUp ticketing off.** Routing is both-and, not
+  either-or — see the escalation-cron doc for the `elseif` bug this used to be.
+- As of 2026-08-04 prod `Logs.IssueEmailAddress` is empty, so this is the **first business-routed
+  issue on the platform**: `processBusinessIssue()` will execute in production for the first time
+  when this issue next fires. Watch that run rather than assuming it works.
+
+## Change history
+- 2026-08-04 — Documented that MITS rejection alerting moved (worker `5521854b`, 2026-08-03) from a
+  hardcoded in-cron `App_Email_Agilant` to `App_Exception_Business` +
+  `Logs.IssueEmailAddress` recipients, so adding an alert recipient is a data change, not a code
+  change. Recorded the deliberate removal of the PII-bearing payload dump from the alert, the
+  look-up-by-`issueKey` / `clientId = 0` / `NOT EXISTS` migration pattern, and that recipient
+  configuration only takes effect on the next live occurrence. (dfranks)
+</content>
+</invoke>

package/knowledge/clients/compass-usa/profile.md

@@ -16,10 +16,11 @@ project: _Underscore
 client: compass-usa
 type: profile
 status: active
-updated: 2026-07-16
-owners: [jcardinal, bala, tcox, apeterson]
+updated: 2026-08-04
+owners: [jcardinal, bala, tcox, apeterson, dfranks]
 files: []
 related:
+  - features/mits-sales-order-transmission-alerting.md
   - features/asn-to-item-fulfillment.md
   - features/cost-centers.md
   - features/approval-decision-flow.md
@@ -75,6 +76,10 @@ separate, related client (see its own profile).
   `worker/crons/toga2/compass/workflow/3b_import_strategic_systems_advance_shipping_notices.php`.
 
 ## Key features (this client)
+- [MITS Sales-Order Transmission — Rejection Alerting](features/mits-sales-order-transmission-alerting.md)
+  — MITS rejection alerts are `App_Exception_Business` Issues with recipients in
+  `Logs.IssueEmailAddress`, **not** an email in the cron: adding an alert recipient is a
+  `dbchanges2/Logs/` data change with no PHP.
 - [Compass ASN → ItemFulfillment Auto-Creation](features/asn-to-item-fulfillment.md) — the
   `_Model_Compass_AdvanceShippingNotice::postPost` handler.
 - [Cost Centers (Unit Locations, numeric-only)](features/cost-centers.md) — what a Compass

package/knowledge/standalone/apps/claude/INDEX.md

@@ -5,3 +5,4 @@
 | [Harness Distribution — How a knowledge.js Fix Reaches Teammates](workflows/harness-distribution.md) | There are **two copies** of `knowledge.js` on a developer's machine, and the skills use only one of them. | claude/scripts/install.js, claude/knowledge.js |
 | [Knowledge Base Publish / Push Pipeline](workflows/knowledge-publish-pipeline.md) | How `/capture` and `/session-save` get knowledge docs from your working tree into `agilantsolutions/claude` on `_main`. | claude/knowledge.js |
 | [MCP Tool Usage — "Talos" Means the Internal Knowledge Base, and Run Your Own SELECTs](workflows/mcp-tool-usage.md) | Two MCP connections are available in every session, and **Claude is expected to reach for them on its own judgement — nobody should have to say the word "MCP".* | claude/.claude/skills/kickoff/SKILL.md, claude/.claude/skills/capture/SKILL.md |
+| [Ticket-Branch Creation Safety — How /work-ticket Armed Pushes to Production Branches](workflows/ticket-branch-creation-safety.md) | For roughly a month, `/work-ticket` created ticket branches whose **git upstream was the production branch**. | claude/skills/work-ticket/SKILL.md |

package/knowledge/standalone/apps/claude/workflows/ticket-branch-creation-safety.md

@@ -0,0 +1,91 @@
+---
+title: Ticket-Branch Creation Safety — How /work-ticket Armed Pushes to Production Branches
+framework: "standalone"
+repo: claude
+project: Claude Harness
+client: shared
+type: workflow
+status: active
+updated: 2026-08-04
+owners: ["dfranks"]
+files:
+  - claude/skills/work-ticket/SKILL.md
+related:
+  - ./harness-distribution.md
+  - ../../../2.0/apps/worker2/workflows/ticket-to-pseudocode-planning.md
+---
+
+## Summary
+
+For roughly a month, `/work-ticket` created ticket branches whose **git upstream was the
+production branch**. Claude's own pushes were unaffected, but a developer using VS Code's
+**Commit + Sync** mid-session pushed ticket code straight to `_main`/`_production` — no PR, no
+review. Two such pushes landed in two days before it was found.
+
+The mechanism is generic git behavior, so it is worth understanding beyond this one skill.
+
+## Root cause
+
+Branch creation was `git checkout -b TRUE-XXXX "origin/$DEFAULT"`. Git's
+**`branch.autoSetupMerge` defaults to `true`**, so branching from a *remote-tracking ref* adopts
+that ref as the new branch's **upstream**. The ticket branch therefore tracked
+`origin/_main` / `origin/_production` until the skill's later `git push -u` re-pointed it.
+
+**Trigger asymmetry is why it hid.** Claude pushes with `git push -u origin <TICKET>` — an
+explicit refspec, which ignores the upstream and was always safe. VS Code's Commit + Sync resolves
+the destination **from the upstream**, so any developer commit made before the skill's first push
+went to the production branch.
+
+**A safety fix created the exposure.** Commit `6d155f7` (2026-07-09) moved branch creation from
+*after* implementation to *before* any code is written — itself a genuine fix, because a stale
+local `_production` 12 commits behind nearly reverted upstream work. But it turned a seconds-long
+armed window into the **entire implementation phase**. Correctness fixes can lengthen an existing
+window; check what else depends on the timing you moved.
+
+## The fixes
+
+1. **Skill:** branch creation uses **`--no-track`** *and* publishes the empty branch immediately
+   (`git push -u origin TRUE-XXXX`), so the branch has a correct upstream before any code exists.
+   The later push is a plain `git push origin <TICKET>`.
+2. **Global git config:** `branch.autoSetupMerge=simple` (a checkout can no longer adopt a
+   production upstream, even outside the skill) plus `push.autoSetupRemote=true`.
+3. **`pre-push` hook** in each app repo rejecting
+   `refs/heads/(_main|_production|_beta|_alpha|master)`, override `ALLOW_PROD_PUSH=1`. This also
+   catches Claude, because Claude shells out to real `git`.
+4. **Deny rules** in `.claude/settings.json`.
+5. Existing branches still tracking a production branch were **disarmed** individually.
+
+`rework-ticket` was never affected — it checks out an existing remote branch.
+
+## Recovery: getting an unreviewed commit off an unprotected default branch
+
+Do **not** force-push. Instead:
+
+1. **Revert via PR** on the default branch.
+2. **Rebuild the ticket branch as a NEW commit** on the post-revert base.
+
+> **The revert-remerge trap:** if you keep the original commit and merge that branch later, the
+> original commit is already an **ancestor** of the default branch, so the merge is a silent no-op
+> and the change **never comes back**. The revert wins and nothing warns you.
+
+## Gotchas / known issues
+
+- **Everything installed here is local and does not survive a fresh clone.** Hooks and git config
+  are per-machine; a teammate or a new laptop is unprotected until they redo it.
+- **Standing exposure (not fixed):** `dbchanges2/_main`, `worker/_production`,
+  `worker2/_production` and `api2/_production` have **no branch protection and no rulesets**.
+  Server-side protection is the only durable fix and needs an admin GitHub account.
+- **Diagnosing this class of incident:** `git reflog` on the remote-tracking branch shows
+  `update by push` entries — that is how both incidents were confirmed and how the other four
+  repos were cleared. Pre-existing `_beta` pushes are normal manual practice; check dates before
+  calling something an incident.
+- Auditing "did this ever happen elsewhere?" requires sweeping **every** repo the skill touches,
+  not just the one that surfaced the bug.
+
+## Change history
+- 2026-08-04 — Documented the `branch.autoSetupMerge` root cause, the Claude-vs-VS-Code trigger
+  asymmetry that hid it for a month, and the fixes (`--no-track` + immediate branch publish,
+  global `autoSetupMerge=simple`, per-repo `pre-push` hook, settings deny rules). Recorded the
+  revert-then-rebuild recovery pattern and its revert-remerge no-op trap, plus the standing lack
+  of server-side branch protection on four app repos. (dfranks)
+</content>

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "toga-ai",
-  "version": "1.0.511",
+  "version": "1.0.512",
   "description": "TOGA Technology Team Claude Knowledge System — shared AI coding harness with skills, knowledge base CLI, and project installer for Claude Code.",
   "keywords": [
     "claude",