# Pass 1 — Classification

Date: 2026-05-16
Source: `Z:\public_html\todos\todos.json` (schema v3, 215 tasks total; 110 pending, 1 in_progress, 101 completed, 3 cancelled)
Last human review: 2026-05-12 23:09 ET (4 days stale)
Last nightly run: 2026-05-15 23:05 ET

## Bucket tally

| Bucket | Count |
|---|---|
| STILL_PENDING_REAL | 91 |
| EXECUTION_READY | 16 |
| PROBABLY_SHIPPED | 0 |
| VAGUE_CANDIDATE | 2 |
| SUPERSEDED | 1 |
| PROBLEM_RESOLVED_BY_DIFFERENT_ARCHITECTURE | 0 |
| STALE_RELEVANT | 0 (folded into STILL_PENDING_REAL with `stale_no_note_update` entropy flag) |

**Total = 110**, matching pending count.

> *Updated 2026-05-16 consistency pass: 5 tasks were initially mis-bucketed
> as PROBABLY_SHIPPED by a regex misfire. Re-bucketed to STILL_PENDING_REAL after
> residual-check; 2 of the 5 then re-qualified as EXECUTION_READY. See PROBABLY_SHIPPED
> section below and `audit_consistency_check.md`.*

The in_progress task `mgp-drop-vital-independent-source` (claude / P0 / mgp) is excluded from classification — it's the active focus per `meta.today_focus[0]`.

## EXECUTION_READY (16)

All meet: owner=claude, P0/P1, no architectural decision required, no upstream blocker in notes, not subject to HUD-off-limits.

| id | priority | property | rev | unlocks | compounding | supervised |
|---|---|---|---|---|---|---|
| p0-freshness-dashboard | P0 | foundation | HIGH | 1 | YES (7 props) | no |
| p0-orchestrator | P0 | foundation | MEDIUM | 0 | YES (5 props) | no |
| p0-identity-fn | P0 | foundation | MEDIUM | 3 | YES (4 props) | no |
| p0-param-surface-registry | P0 | param_surface | MEDIUM | 2 | YES (3 props) | no |
| data-freshness-audit | P0 | foundation | HIGH | 2 | no | no |
| catalog-refactor-shared-json | P1 | foundation | LOW | 4 | YES (5 props) | **YES** |
| add-user-role-approval | P1 | foundation | LOW | 4 | no | **YES** |
| refactor-scientific-overlay-js-to-load-from-system | P1 | scorecards | MEDIUM | 2 | no | no |
| strategy-vpi-rules | P1 | scorecards | MEDIUM | 3 | no | no |
| mgp-daily-stop-pipe | P1 | daily_stop | MEDIUM | 0 | no | no |
| pm-nightly-task | P1 | pm_system | MEDIUM | 0 | no | no |
| pm-nightly-silent-failure-detection | P1 | pm_system | HIGH | 0 | no | no |
| scorecard-time-of-day-equity | P1 | scorecards | HIGH | 0 | no | no |
| scorecard-strategy-time-of-day | P1 | scorecards | MEDIUM | 0 | no | no |
| ux-cross-nav | P1 | foundation | MEDIUM | 0 | no | no |
| mtf-vp-tradingview-handler | P1 | mtf_vp_scanner | MEDIUM | 0 | no | no |

Confidence on bucket = **medium** for all. Reasoning: each task has a non-trivial note + a concrete property + a current claude owner, but none has an explicit done_when yet — Pass 4 adds those.

## SUPERSEDED (1)

| id | superseded_by | confidence |
|---|---|---|
| mgp-recap-stability | mgp-cron-skipped-5-14 | **high** |

Rationale: `mgp-recap-stability` (created 2026-05-01) tracked 3 recap silent failures. `mgp-cron-skipped-5-14` (created 2026-05-15) is broader — covers 3 fire slots over 2 trading days — and was nominated to today_focus[2]. The older task adds no information the newer one lacks.

## PROBABLY_SHIPPED (0 — after residual check)

The analyzer's regex initially bucketed 5 tasks as PROBABLY_SHIPPED because their notes mentioned "shipped" / "already" / "Path A" in past tense alongside the pending work. Residual-check on each found real remaining scope. All 5 were re-bucketed.

| id | initial bucket | reason for misfire | final bucket | confidence |
|---|---|---|---|---|
| mgp-conservative-moderate-aggressive | PROBABLY_SHIPPED | "Matches `entry_type` *already in use*" refers to the column, not the task | STILL_PENDING_REAL | high |
| mgp-key-level-enrichment | PROBABLY_SHIPPED | "`indicator_alerts` *already has* VWAP bands..." describes the schema, not task progress | STILL_PENDING_REAL | high |
| ux-cross-nav | PROBABLY_SHIPPED | `/mgp/` already has the top-nav, but the task is to extend it to 7 other properties | **EXECUTION_READY** (re-qualified) | medium |
| mtf-vp-tradingview-handler | PROBABLY_SHIPPED | Path A shipped 4/28; Path B is the actual task and is not stubbed beyond the endpoint | **EXECUTION_READY** (re-qualified) | medium |
| acct-v2-fills-poller-bridge | PROBABLY_SHIPPED | v1 fills *already arrive* into the table, but the dual-write shim is the actual task | STILL_PENDING_REAL | high |

After the residual check, the **actual PROBABLY_SHIPPED count is 0**. Two of the five (`ux-cross-nav`, `mtf-vp-tradingview-handler`) re-qualified as EXECUTION_READY because they're claude-owned + P1 + no upstream blocker; the other three either lack a priority or have notes that describe schema state rather than task progress.

## VAGUE_CANDIDATE (2)

Empty note + no path + no clear done_when. These need either reframe-with-DoD or archive.

| id | suggested action |
|---|---|
| remove-na-labels | reframe: depends on `wire-webhooks` shipping; DoD = "scorecards no longer show 'webhook-data N/A' anywhere for trades that have an indicator_alert row". Drop empty placeholder. |
| strategy-backside-backtest | reframe: depends on `strategy-backside-rules`; DoD = "/backtests/backside-scalp/ index renders walk-forward stats". |

Neither rises to archive; both have clear intent, just no DoD.

## STILL_PENDING_REAL (88)

The bulk. Subdivided informally:

- **Foundation phase-0 P0 deliverables not in EXECUTION_READY (2):** p0-daily-summary-view (blocked on user-owned `run-scientific-layer-migrations`), p0-indicator-alerts-ddl (user-owned, P0)
- **acct-v2 BUGS reported by James (2):** acct-v2-bug-risk-bump-ignores-date-range (P0 current_priority, owner=jimbo), acct-v2-bug-ticket-meter-ignores-date-range (P0, owner=jimbo)
- **MGP cutover-adjacent (3):** mgp-cron-skipped-5-14 (P0, owner=James with capitalization drift), mgp-historical-pages-cleanup (P0, owner=james — compliance), and the in_progress `mgp-drop-vital-independent-source` itself
- **HUD-off-limits supervised work (4):** acct-v2-hud-trader-state-card, acct-v2-hud-composure-trend-card, acct-v2-hud-strategies-mood-card, wire-live-grade-column-into-accountability-hud-v2-
- **Architecturally-blocked clusters:** strategy_library tasks gated on add-user-role-approval (4), phase-4 multi-trader tasks gated on multi-trader-scorecards (4), strategy-backtest tasks gated on rules being codified (3)
- **risk_hud cluster (5):** all blocked on the dedup decision `risk-hud-accountability-hud-dedup-review` (P1, user-owned)
- **user-owned blockers (~14):** journal-user-steps (P0), run-scientific-layer-migrations (P0), p0-indicator-alerts-ddl (P0), pm-activity-scanner-deploy (P1), claude-code-webdav-path-workaround (P1), acct-v2-kore-tick-bridge-forward-test (P1), trade-action-requests-feature (P1), promote-dev-to-prod (P1), recent-migrations-doc-delta (P2), rescue-localstorage (none), verify-enrich-on-server (P1), cron-update-freshness (P1), roberto-onboarding-from-artifact (P1), phase-4-trader-of-the-week (P1)
- **Other-owner work:** alex (3), roberto (2), jimbo (9 — all acct-v2)

## Confidence summary

- High confidence: bucket assignment is well-grounded for **18 tasks** — the 14 EXECUTION_READY + SUPERSEDED + user-blocked-by-explicit-note (4).
- Medium confidence: the **78 STILL_PENDING_REAL** with named architectural blockers, plus the **2 VAGUE_CANDIDATE**.
- Low confidence: **0 tasks** — every pending task has either a note, a path, or an owner that supports a bucket call.

## Methodology notes

- "EXECUTION_READY" requires the task to be **safe to attempt with no live human supervision**. The reversibility classification in Pass 3 is what determines whether it actually belongs in the autonomous queue (HIGH/MEDIUM) or the supervised queue (LOW).
- "PROBABLY_SHIPPED" candidates always got a residual-check before final assignment. Notes that say "STATUS 2026-05-06: scaffolded, outstanding..." indicate real remaining work and do NOT qualify.
- HUD-off-limits memory (`feedback_hud_off_limits.md`) gates any task whose property is `accountability_hud` AND whose action involves editing the live cockpit. Diagnostic + bug-tracking tasks on the same property are not gated.
