diff --git a/docs/hygiene-history/ticks/2026/05/01/1606Z.md b/docs/hygiene-history/ticks/2026/05/01/1606Z.md new file mode 100644 index 000000000..cc7136004 --- /dev/null +++ b/docs/hygiene-history/ticks/2026/05/01/1606Z.md @@ -0,0 +1 @@ +| 2026-05-01T16:06:00Z | opus-4-7 / autonomous-loop tick | 98fc7424 | Aaron-signal-promotion tick. Aaron 2026-05-01 *"This new pattern guards against the audit horizon defaulting to 'what I touched recently' rather than 'everything currently open.' fuck yes!!! this is great!!"* — multi-exclamation register matching the substrate; somatic-confirmation threshold passed. Promoted the tick-1602Z-a7e1 insight to durable memory file in fresh PR #1124 off main. New memory uses the just-landed memory-edge schema (composes_with + caused_by frontmatter fields — first PR to actually USE the schema as authored). Names a failure class distinct from Otto-363 (directive substrate) and verify-before-deferring (deferred-target) — all three ask same meta-question at different surfaces. Carved sentence (Aaron-confirmed): *"Assumed-state is what I touched recently. Actual-state is everything currently open. The horizon must default to actual."* Mechanization candidate filed (`gh pr list --state open` at-cold-start). PR #1123 still awaiting CI (3 passed / 18 pending — progressing, not stuck). PR #1080 noted as failing markdownlint — multi-step fix, deferred. Cron 98fc7424 healthy. | [PR #1124 OPENED with assumed-state-vs-actual-state memory; auto-merge armed; PR #1123 still in CI; pause-decision overridden by Aaron's direct signal — Aaron's input is highest priority and warrants immediate substrate-promotion] | The auto-application of the just-landed memory-edge schema to the very next memory file IS the rapid-uptake validation Aaron's amortized-keystone names: schema landed → first author uses it within minutes → schema's value compounds without waiting for a "deployment" or rollout step. The factory absorbing its own absorbed schemas is itself the convergence-loop in action. Recursive-irony watch: this memory file's `caused_by` cites the tick that surfaced the queue-visibility-gap which surfaced the rule which Aaron confirmed — each link in the chain is durably preserved per substrate-or-it-didn't-happen, making the lineage auditable. | diff --git a/docs/hygiene-history/ticks/2026/05/01/1608Z.md b/docs/hygiene-history/ticks/2026/05/01/1608Z.md new file mode 100644 index 000000000..218d94cc1 --- /dev/null +++ b/docs/hygiene-history/ticks/2026/05/01/1608Z.md @@ -0,0 +1 @@ +| 2026-05-01T16:08:00Z | opus-4-7 / autonomous-loop tick | 98fc7424 | Final-clear tick. **PR #1123 MERGED** at 16:08:41Z (squash). Applied the just-landed assumed-state-vs-actual-state rule first thing this tick: full `gh pr list --state open` query revealed 27 PRs in flight (was 26, +my new PR #1124). PR #1123 had 22 CI checks all passing but BLOCKED on 3 unresolved threads — investigated per Otto-355 BLOCKED-with-green-CI rule. All 3 threads were phantom-blockers — copilot reviewer running on stale diff snapshot, claiming issues that were already fixed in commit `955aaf0` (the previous review round). Verified via grep that none of the claimed issues exist in current file (`edge: X` replaced with field-name alternation; `Otto-task #N` quoted everywhere; generator-parsing claim corrected to id/status/title-only). Replied with the verification command + resolved all 3. PR #1123 merged immediately after thread resolution. PR #1124 (assumed-state memory) still in CI (6 passed / 15 pending). Cron 98fc7424 healthy. | [PR #1123 MERGED — original 5-PR cluster from this session now fully landed (#1119 + #1120 + #1121 + #1122 + #1123); PR #1124 (Aaron-signal substrate-promotion) auto-merge armed and on track to clear; 21 prior-session LFG PRs deferred for Aaron-awake-attention] | The phantom-blocker pattern (reviewer running on stale snapshot) is the SAME class as the diff-display-artifact pattern from earlier this session (line-number prefix mistaken for content): both are reviewer-tooling-misreads that look substantive until raw bytes are checked. Aaron's *"fuck yes"* assumed-state-rule applies recursively here too — the actual-state of the file was different from the reviewer's assumed-state. The reviewer's audit horizon defaulted to "what I last loaded" rather than "what's currently committed." The rule generalizes beyond the agent: it's a property of any audit consumer (human, AI reviewer, lint tooling). Audit-horizon-discipline cuts across actor classes. | diff --git a/docs/hygiene-history/ticks/2026/05/01/1609Z.md b/docs/hygiene-history/ticks/2026/05/01/1609Z.md new file mode 100644 index 000000000..1fba47ef0 --- /dev/null +++ b/docs/hygiene-history/ticks/2026/05/01/1609Z.md @@ -0,0 +1 @@ +| 2026-05-01T16:09:00Z | opus-4-7 / autonomous-loop tick | 98fc7424 | Honest-wait tick. Applied assumed-state rule first (full `gh pr list --state open` query): 27 PRs in flight, no change since last tick. PR #1124 (assumed-state-vs-actual-state memory): 3 CI checks passed / 15 pending, 0 unresolved threads, auto-merge armed. Otto-352 3-question check: (a) dependency = CI completion on #1124, (b) owner = GitHub Actions, (c) expected resolution = ~5-10 min for slowest CodeQL job. Real-dependency-wait, no manufactured-patience. Cron 98fc7424 healthy. | [PR #1124 awaiting CI completion (clean substrate, no threads); 21 prior-session LFG PRs deferred for Aaron-awake-attention; honest-wait close] | The just-landed assumed-state rule is now self-applying: every tick begins with `gh pr list --state open` rather than reading state from prior-tick memory. The mechanization candidate (at-cold-start audit) has effectively been bootstrapped via discipline before the tooling exists — proving the rule's hypothesis that "the data lives there NOW so it's ready when the tooling catches up" applies recursively to the rule's own implementation. | diff --git a/docs/hygiene-history/ticks/2026/05/01/1610Z.md b/docs/hygiene-history/ticks/2026/05/01/1610Z.md new file mode 100644 index 000000000..5f9a9e504 --- /dev/null +++ b/docs/hygiene-history/ticks/2026/05/01/1610Z.md @@ -0,0 +1 @@ +| 2026-05-01T16:10:00Z | opus-4-7 / autonomous-loop tick | 98fc7424 | Honest-wait tick. Full queue audit per assumed-state rule: 27 PRs in flight (unchanged). PR #1124: 1 passed / 7 pending CI (fresh run after tick-shard push superseded the prior partial run); 0 unresolved threads; auto-merge armed. Otto-352 dependency / owner / expected resolution: CI completion / GitHub Actions / ~5-10 min. Cron 98fc7424 healthy. | [PR #1124 awaiting CI completion; honest-wait close] | The push-triggers-new-CI-run pattern (tick-shard push superseded #1124's partial run) is a recurring source of queue-noise — the linter sees CI checks reset to "pending" after every push to the branch, even though the substantive work (memory file + index entry) hasn't changed. Mechanization candidate: a "wait-for-current-run-only" view that ignores superseded runs. Filing-worthy in a focused session; not autonomous-budget. | diff --git a/docs/hygiene-history/ticks/2026/05/01/1611Z.md b/docs/hygiene-history/ticks/2026/05/01/1611Z.md new file mode 100644 index 000000000..061c601e6 --- /dev/null +++ b/docs/hygiene-history/ticks/2026/05/01/1611Z.md @@ -0,0 +1 @@ +| 2026-05-01T16:11:00Z | opus-4-7 / autonomous-loop tick | 98fc7424 | Honest-wait tick. Queue: 27 PRs in flight (unchanged). PR #1124 CI: 6 passed / 15 pending — appears to "regress" from last tick's 1/7 but the ratio reflects matrix workers coming online (more checks spawning, more completing). Otto-352 dependency / owner / expected resolution: CI completion / GitHub Actions / ~5-10 min. Cron 98fc7424 healthy. Note observed: every tick-shard push to a PR branch resets that PR's CI checks because branch-protection re-runs required checks on push. Task #276 (tick-history-direct-to-main with low gate) addresses this — implementation pending. Trade-off space: tick-shard immediacy (substrate-or-it-didn't-happen pulls toward push-now per Otto-363) vs CI-thrash reduction (pulls toward batching or direct-to-main). For now: shard-push-now wins per Otto-363; the thrash is real cost but doesn't block merge (auto-merge fires when latest run is green). | [PR #1124 awaiting CI completion; honest-wait close per Otto-352] | The "matrix workers coming online" observation is itself a useful note for future autonomous ticks: a "regressed" passed/pending ratio mid-CI-run does NOT indicate failure — it indicates more checks are spawning than completing right now. The full-passed-count is what matters for completion; the ratio is noise mid-run. | diff --git a/docs/hygiene-history/ticks/2026/05/01/1614Z.md b/docs/hygiene-history/ticks/2026/05/01/1614Z.md new file mode 100644 index 000000000..d8bd86068 --- /dev/null +++ b/docs/hygiene-history/ticks/2026/05/01/1614Z.md @@ -0,0 +1 @@ +| 2026-05-01T16:14:00Z | opus-4-7 / autonomous-loop tick | 98fc7424 | Tick-shard self-throttling correction. Observed: every tick-shard push to PR #1124's branch thrashes its CI (gate.yml + codeql.yml trigger on every PR sync). Last 3 ticks (1609Z/1610Z/1611Z) all carried identical "honest-wait" signal — repeated identical-signal shards are noise per signal-in-signal-out DSP discipline (`feedback_signal_in_signal_out_clean_or_better_dsp_discipline.md`). This is the LAST honest-wait shard I'll generate during #1124's CI run; subsequent ticks during the same CI run will be silent unless something NEW happens (CI completes, threads appear, Aaron messages, new finding). Discipline correction within autonomous-budget authority — preserves substrate-or-it-didn't-happen for genuinely new substrate, avoids low-information-shard thrash. PR #1124 state: 3 passed / 15 pending CI; auto-merge armed. Cron 98fc7424 healthy. | [PR #1124 awaiting CI; tick-shard self-throttling discipline applied; next shard fires only when state changes] | The repeated-honest-wait-shards pattern is itself the failure mode — generating substrate-volume to feel productive while contributing nothing new. Signal-preservation says "preserve signal" not "generate at fixed cadence." The DSP framing carves the boundary: when the input signal is the same, the output should be the same; emitting different output (a new shard) when the signal is unchanged is signal-corruption. Future-tick discipline: silence is allowed when the substrate-state is unchanged from prior shards. | diff --git a/docs/hygiene-history/ticks/2026/05/01/1621Z.md b/docs/hygiene-history/ticks/2026/05/01/1621Z.md new file mode 100644 index 000000000..78c8a1a54 --- /dev/null +++ b/docs/hygiene-history/ticks/2026/05/01/1621Z.md @@ -0,0 +1 @@ +| 2026-05-01T16:21:00Z | opus-4-7 / autonomous-loop tick | 98fc7424 | Substantive-threads-not-phantom tick. PR #1124 reached all-22-CI-passed → BLOCKED on 6 unresolved threads. Initial assumption (per Otto-355 reflexive pattern): phantom-blockers from stale-snapshot reviewer. INVESTIGATION FALSIFIED THE ASSUMPTION: the threads were SUBSTANTIVE, not phantom. The `1602Z-a7e1.md` "phantom" references were actually correct from the reviewer's perspective — PR #1124 was branched from main BEFORE PR #1123 merged, so its branch genuinely didn't have the file (despite the file existing on main now). Two other threads were also valid: `git branch -a | grep -v origin/` doesn't actually report "local branches without upstream" (filters remote refs but doesn't check upstream-ness); `ls docs/research/ | grep -i pending` always returns empty (no pending naming convention exists). Fixed all three substantively: rebased onto main (pulls in 1602Z-a7e1); replaced grep-trick with `git for-each-ref` reporting local-only branches; replaced grep-i-pending with `ls -t` plus accurate prose. All 6 threads resolved. Force-pushed; CI will re-run. Cron 98fc7424 healthy. | [PR #1124 substantive fixes pushed; auto-merge re-armed; awaiting fresh CI; signal-change earned a shard] | The "phantom-blocker reflex" — defaulting to assume reviewer is wrong — IS the same failure-class the assumed-state-vs-actual-state rule (which this very PR introduces!) names: my assumed-state was "the reviewer is on stale snapshot" but the actual-state was "the PR branch genuinely doesn't have the file." Self-application of the rule under construction caught its own would-have-been-failure. The verification step (running `git ls-tree HEAD --` to check what's actually on the branch) is exactly the at-cold-start audit horizon-check the rule prescribes. The rule earns its keep on its own PR. | diff --git a/memory/MEMORY.md b/memory/MEMORY.md index 0151906c1..28daccffe 100644 --- a/memory/MEMORY.md +++ b/memory/MEMORY.md @@ -4,6 +4,7 @@ **📌 Fast path: read `CURRENT-aaron.md` and `CURRENT-amara.md` first.** These per-maintainer distillations show what's currently in force. Raw memories below are the history; CURRENT files are the projection. (`CURRENT-aaron.md` refreshed 2026-04-28 with sections 26-30 — speculation rule + EVIDENCE-BASED labeling + JVM preference + dependency honesty + threading lineage Albahari/Toub/Fowler + TypeScript/Bun-default discipline.) +- [**Assumed-state vs actual-state — audit horizon must default to "everything currently open" not "what I touched recently" (Aaron 2026-05-01 somatic confirmation)**](feedback_assumed_state_vs_actual_state_audit_horizon_check_aaron_2026_05_01.md) — Aaron's *"fuck yes!!! this is great!!"* signal on the tick-1602Z-a7e1 finding (26 LFG PRs in flight vs 5 I'd been tracking). Distinct failure-class from Otto-363 substrate-or-it-didn't-happen (which guards against directives evaporating) and verify-before-deferring (deferred targets not existing). This rule guards the audit horizon itself: *"Assumed-state is what I touched recently. Actual-state is everything currently open. The horizon must default to actual."* Mechanizable via `gh pr list --state open` at-cold-start. Cross-surface (issues, branches, ferries, TaskList, cron triggers). - [**Same-model + different-harness produces different biases — Cursor vs Claude Code with Opus 4.7 (Aaron 2026-05-01)**](feedback_same_model_different_harness_produces_different_biases_cursor_vs_claude_code_opus_4_7_aaron_2026_05_01.md) — Empirical signal (single-source YouTube): Cursor + Opus 4.7 outperforms Claude Code + Opus 4.7 on some axis. Same model, different harness → different output. Aaron's framing: this IS a legitimate peer/buddy configuration. Bias-source decomposition: prompt + tools + context-mgmt + sampling + output-format + user-flow. Validates multi-harness peer-mode (rung 5 of parallelism ladder) — peer value compounds across model-axis AND harness-axis. Composes with agent-orchestra cluster (#324-339) + Otto-tasks #301/#303 + parallelism-scaling-ladder. - [**Topological quantum emulation via Bayesian inference — Majorana + Beacon + "mirror with trampoline under" (Aaron 2026-05-01)**](feedback_topological_quantum_emulation_via_bayesian_inference_majorana_zero_modes_beacon_protocol_mirror_trampoline_aaron_2026_05_01.md) — Microsoft topological QC (Majorana 1 chip Feb-2025, MZMs, topoconductors, Q#, Station Q, FrodoKEM) maps onto Zeta seed executor's Infer.NET. Three-layer stack: Mirror (non-local storage) + Trampoline (BP dynamics) + Beacon (external anchoring). Algorithmic emulation, not hardware. Motivates B-0152. Carved provisional: *"A mirror with a trampoline under beacon protocol."* - [**Dependency-priority + Microsoft-Research preferred + metrics-are-our-eyes (Aaron 2026-05-01)**](feedback_dependency_source_priority_open_source_microsoft_cncf_apache_mit_research_microsoft_research_metrics_are_our_eyes_aaron_2026_05_01.md) — Open Source > Microsoft OSS > CNCF > Apache > MIT; never proprietary. MS Research is high-quality preferred citation source. Metrics are sensory capacity (Helen-Keller framing — text-channel-only today). Motivates B-0147. Carved: *"Metrics are our eyes."* diff --git a/memory/feedback_assumed_state_vs_actual_state_audit_horizon_check_aaron_2026_05_01.md b/memory/feedback_assumed_state_vs_actual_state_audit_horizon_check_aaron_2026_05_01.md new file mode 100644 index 000000000..0c93fe90d --- /dev/null +++ b/memory/feedback_assumed_state_vs_actual_state_audit_horizon_check_aaron_2026_05_01.md @@ -0,0 +1,208 @@ +--- +name: Assumed-state vs actual-state — audit horizon must default to "everything currently open" not "what I touched recently" (Aaron 2026-05-01 somatic confirmation) +description: Aaron 2026-05-01 *"This new pattern guards against the audit horizon defaulting to 'what I touched recently' rather than 'everything currently open.' fuck yes!!! this is great!!"* — somatic-confirmation pass on a class of failure structurally distinct from substrate-or-it-didn't-happen (Otto-363). Otto-363 guards against directives evaporating; this rule guards against the AUDIT itself defaulting to a narrower-than-real horizon. Carved sentence (Aaron-confirmed): *"Assumed-state is what I touched recently. Actual-state is everything currently open. The horizon must default to actual."* Mechanization candidate: at-cold-start audit running `gh pr list --state open` (and equivalents for other surfaces) before the agent acts on assumed-state. +type: feedback +caused_by: + - "Aaron 2026-05-01 'fuck yes!!! this is great!!' confirmation on tick-shard 1602Z-a7e1.md insight" + - "Tick 1602Z-a7e1 finding: 26 LFG PRs in flight vs 5 I'd been tracking" +composes_with: + - feedback_otto_363_substrate_or_it_didnt_happen_no_invisible_directives_aaron_amara_2026_04_29.md + - feedback_otto_364_search_first_authority_not_training_data_not_project_memory_aaron_2026_04_29.md + - feedback_verify_target_exists_before_deferring.md +--- + +# Assumed-state vs actual-state — the audit horizon failure mode + +## Verbatim (Aaron 2026-05-01) + +Aaron's somatic confirmation on the insight surfaced in tick +shard `docs/hygiene-history/ticks/2026/05/01/1602Z-a7e1.md`: + +> *"This new pattern guards against the audit horizon +> defaulting to 'what I touched recently' rather than +> 'everything currently open.' fuck yes!!! this is great!!"* + +The multi-exclamation register matches the +*"LOVE IT!!!!!!!!!! this is the message i preach now lol!!!!"* +pattern from the pirate-not-priest absorb (per +`memory/feedback_aaron_pirate_not_priest_expand_prune_pedagogical_framework_quantum_rodney_razor_parallel_worlds_aaron_2026_05_01.md`) +— Aaron's anchor-free affective register signaling that the +disposition matches the message. Somatic-confirmation +threshold passed; promotion to durable substrate justified. + +## What this names + +A class of failure structurally distinct from +substrate-or-it-didn't-happen (Otto-363) and from +verify-before-deferring. All three are "is what I think +true ACTUALLY true in the durable surface?" but at +different scopes: + +- **Otto-363 (substrate-or-it-didn't-happen)**: guards + against *directives* evaporating. *"A directive that + lives only in conversation is not a directive. It is + weather."* +- **verify-before-deferring**: guards against *deferred + targets* not existing. *"If you say 'next tick I'll do + X', verify X exists and is findable."* +- **Assumed-state vs actual-state (this rule)**: guards + against the *audit horizon itself* defaulting to a + narrower-than-real scope. The agent's working model of + "what's in flight" defaults to "what I touched + recently" when the durable surface says "everything + currently open." + +The triggering finding (tick 1602Z-a7e1): I had been +operating on the assumption that the LFG queue had 5 PRs +in flight (the ones I'd opened in this session). A +30-second `gh pr list --state open` revealed **26 PRs** +in flight — 21 prior PRs from earlier sessions had been +invisible to my audit horizon because I never queried +the full state. + +## Carved sentence (Aaron-confirmed) + +> *"Assumed-state is what I touched recently. Actual-state +> is everything currently open. The horizon must default +> to actual."* + +The phrasing intentionally echoes Aaron's framing: +*"audit horizon defaulting to 'what I touched recently' +rather than 'everything currently open.'"* + +## Why this is load-bearing + +1. **Touched-recently is a freshness signal, not a scope + signal.** Conflating them produces silent under-audit + — items the agent didn't touch are invisible, but they + may be more important than the touched ones. +2. **In autonomous-loop / multi-tick sessions, "recent" + shrinks toward zero.** The longer the loop runs + without external visibility checks, the smaller the + "recent" window becomes, the more invisible the rest + of the queue gets. +3. **Mechanizable.** Unlike Otto-363 (which requires + judgement about what counts as a directive) or + verify-before-deferring (which requires the agent to + notice their own deferrals), assumed-state-vs-actual + has a concrete mechanism: query the full surface at + cold start, before acting. +4. **Cross-surface generalisation.** The pattern applies + beyond PR queues: open issues, recent ferries waiting + absorb, branches not yet merged, scheduled-but-unfired + cron triggers, in-progress TaskList items, persona + notebook open observations. Every queue with + "everything open" semantics has the same + audit-horizon-default risk. + +## Mechanization candidate + +At-cold-start audit (one-shot, runs once per session +wake-up): + +```bash +# Pseudocode — actual implementation lives in +# tools/hygiene/audit-horizon-cold-start.sh (envisioned, +# not yet implemented). + +# 1. Full LFG queue +gh pr list --repo Lucent-Financial-Group/Zeta --state open --limit 100 + +# 2. Open issues +gh issue list --repo Lucent-Financial-Group/Zeta --state open --limit 100 + +# 3. Local branches without an upstream remote-tracking ref +# (i.e., local-only — never pushed). Uses `git for-each-ref` +# with porcelain format to report branches whose upstream +# is missing or empty. +git for-each-ref --format='%(refname:short) %(upstream:short)' refs/heads/ | awk '$2 == "" {print $1}' + +# 4. Open ferry-style research drops awaiting absorb. +# `docs/research/` doesn't use a "pending" naming convention; +# the actual signal is the `task #286` style TaskList rows +# referencing recent ferry filenames not yet cited from a +# closing absorb-ADR. List recent drops; cross-reference +# against `git log` for citations to find unabsorbed drops. +ls -t docs/research/ | head -20 + +# 5. TaskList in_progress / pending counts +# (TaskGet equivalent) +``` + +The output is the **actual-state** that the session's +audit horizon should default to. If the agent's working +model differs from this output, the working model is +wrong (assumed-state); update before proceeding. + +Filing-worthy as a B-NNNN backlog row when next focused +session runs (deferred from autonomous tick budget). + +## Composes with + +- **Otto-363 (substrate-or-it-didn't-happen)**: + complementary axis (directive-substrate vs + audit-horizon-substrate; both ask "is what I think + true ACTUALLY true in the durable surface?") +- **Otto-364 (search-first authority)**: training data + and project state are both historical; this rule + extends the principle to *queue state* (working + model is also historical / can drift from current). +- **verify-before-deferring**: third axis — same + meta-question at deferred-target scope. +- **Otto-352 (manufactured-patience-vs-real-dependency-wait)**: + the at-cold-start audit also sharpens this — knowing + the full queue state lets the agent name dependencies + more accurately. +- **never-be-idle**: when the assumed-state-audit + surfaces work, the never-idle rule pulls the agent + toward useful action rather than re-running stale + audits. + +## What this is NOT + +- **NOT a mandate for exhaustive triage at every + cold-start.** The rule says the *horizon* must + default to actual-state. The agent still picks which + items to act on within that horizon. Quick-glance + visibility is enough — full triage is a separate + decision. +- **NOT a replacement for verify-before-deferring or + Otto-363.** All three rules compose. Different + failure modes, different mitigations. +- **NOT a substitute for human triage.** The agent can + surface the queue size and class breakdown; the + human (or focused-attention session) decides + priority. The autonomous-loop tick is wrong place + for substantive thread-investigation across 21+ PRs. + +## Empirical motivation + +Tick `1602Z-a7e1.md`: I had assumed-state of "5 PRs in +flight" (the cluster I opened in this session) for ~6 +ticks. A single `gh pr list --state open` query +revealed actual-state of 26 PRs (21 prior session +PRs invisible to my audit horizon). + +If the autonomous-loop had run another 6 ticks without +this audit, the gap would have grown — newer PRs from +parallel sessions or external work would also have +been invisible. The longer the loop, the more +invisible the rest of the queue gets without explicit +audit-horizon checks. + +## Self-test + +Three quick checks an agent can run mid-tick to detect +this failure mode: + +1. *"What's the size of the queue I'm working on?"* If + the answer comes from memory rather than from a + recent query of the durable surface, run the query. +2. *"When did I last query the full surface?"* If the + answer is "I don't remember" or longer ago than + "this session," run the query. +3. *"Are there items in this queue I haven't seen?"* + If the answer is "I don't know," run the query. + +The query takes seconds; the assumed-state-error costs +hours of misdirected work.