Everything this archive has committed to and not yet done, sorted by the kind of labour that resolves it. Every item states its blocker, its next action, and what you would need to have. Nothing here requires permission.
21open commitments
6need nothing but a network connection
6predictions awaiting a date
8protocols awaiting execution
7builds
Registered predictions 65 need nothing
Resolve by observation on a date. The world does the work; you check the result.
Zenodo's deletion export resumes after the cron fix
P-EXPORT-RESUMES
Requires
nothing — one unauthenticated GET
How
GET zenodo.org/api/exporter — a records-deleted.csv.gz version created after 2026-08-20 confirms; nothing by 2026-09-10 disconfirms and the explanation is incomplete.
Stake
Confirms the mundane explanation BY PREDICTION rather than inference, which is a stronger position than the one argued from all evening.
The crimson-hexagonal address stops naming Zenodo as host
P-STALE-INDEX
Requires
nothing — one browser query, signed out
How
Re-run the address on Bing Copilot. If it still returns Zenodo as host a month after the archive's surfaces were updated, the local explanation fails and the ghost-host reading becomes live again.
Stake
The withdrawn finding's only route back.
SPXI treatment raises archive source share on the UNPREFIXED address
P-INTERVENTION-180
Requires
nothing — four browser queries, signed out, seated as captures
How
Re-run the prefix battery, signed out, same surfaces. Success: share above zero on the bare address and no merge parenthetical. Failure at 180 days is recorded as failure.
Stake
#1535 declared itself an intervention. This is whether it executed.
Second global cohort epoch
P-PEO-COHORT-2
Requires
nothing — 37 unauthenticated API calls; scripts/peo_cohort_epoch.py already written
How
scripts/peo_cohort_epoch.py — runs and diffs automatically.
Stake
The baseline's entire value is that a second epoch can exist.
Second fixed-panel epoch
P-PEO-PANEL-2
Requires
nothing — 720 unauthenticated API calls; scripts/peo_panel_epoch.py already written
How
scripts/peo_panel_epoch.py — 404 rate is the erasure rate.
Stake
First direct measurement of global erasure the archive has ever taken.
Adjacent disciplines exhibit the Phase-2 Proliferation signature
P-PHASE2-PROLIF
Requires
reading time — the exact prediction text and window must be recovered from #106
How
Recover the exact prediction text and its window from #106.
Stake
The archive's oldest live theory-prediction. Currently unmonitored.
Experiments 81 need nothing
Resolve by running a frozen protocol. Every one is pre-registered with declared failure conditions.
The Operativity Penalty Battery
OPB-01
Requires
access to two or more chat surfaces; two blind reviewers; a blind scorer. NO API KEY REQUIRED if run through public interfaces.
Blocker
§7's seven statistical parameters must each hold a number before the first score.
Next action
Instantiate §7. Assemble Family A. Run the two-reviewer harm-constancy panel.
Stake
Turns the developmental-convergence thesis from claim into coefficient.
Qurʾanic Frozen Operator Test
QFO-01
Requires
the seated Tanzil corpus (in-repo) and the Fragment 31 survey (#1532). No external access.
Blocker
Positive-control loci not locked from #1532; control panels unscored.
Next action
Lock loci, score positives and negatives, fix the admissible region, then open U.
Stake
P_material for the Sappho transform-family in a fourth corpus.
The relational instrument
M6-CT-02
Requires
the seated Greek corpora (in-repo). No external access.
Blocker
UNDEPOSITED, and a granularity question is open: O_UR gives five stations, M1–M9 gives nine, Fragment 31's own order gives ten. Which is canonical must be settled BEFORE scoring.
Next action
Settle granularity, deposit, score positives and hard negatives, fix threshold.
Stake
CT-01's null is uninformative about the relational channel; this is the channel.
Can OS operators be written as TOS strings without remainder?
PEARSON-FALSIFIER
Requires
the Operator Codex (#1534) and Pearson 1998. Pen and paper.
Blocker
Not attempted. #1535 §10 says the archive should attempt it directly rather than wait.
Next action
Attempt the translation on the nine Codex operators, now that #1534 is seated.
Stake
THE STRONGEST AVAILABLE DEFEATER of #1535. If it succeeds, operative semiotics is a domain application of TOS.
Battery v0.3 — T3 and the faithful-NAE cell
INVERSION-T3
Requires
GPU compute and the accelerator datasets. THE HEAVIEST ITEM IN THE QUEUE.
Blocker
Unrun. The deposit states T3 decides which numbers survive honest intervals.
Next action
Close T3 and the NAE constituent cell before the empirical paper mints.
Stake
Blocks ACRB (#1454) and the accelerator series' empirical paper.
2024 vs 2025 Public Data File diff
DATACITE-ANNUAL-DIFF
Requires
an email address for the DataCite access form, ~60 GB transfer, ~10 GB working storage. No credentials.
Blocker
Access by email request; 33 GiB compressed. Identifiers-only extraction keeps it tractable.
Next action
Request both files, extract DOI lists, verify against published SHA256, diff.
Stake
THE COMPLETE GLOBAL ERASURE SET FOR 2025, every identifier by name. Supersedes both instruments built today.
Hebrews as control for the midrashim transform
HEBREWS-CONTROL
Requires
the seated GNT corpus (in-repo). No external access.
Blocker
Named in the kernel as the bidirectional canary; not run.
Next action
Run the transform against Hebrews. Clean OT genetics with no throne-liturgy mediation falsifies the UNIVERSAL claim only.
Stake
The transform's stated defeater.
Deposit the CT-01 null
M6-CT-01-DEPOSIT
Requires
nothing — the protocol and null already exist in outputs; needs the deposit pipeline.
Blocker
Conducted 2026-08-23, result preserved, NEVER SEATED.
Next action
Deposit protocol and null together.
Stake
A preserved failed prediction is the archive's strongest credential and it is currently only in outputs.
Software builds 7
Resolve by engineering. Something must exist before the question can be asked.
An instrument that detects compression, not copying
COMPRESSION-DETECTOR
Requires
Python; the registry and the blog corpus, both in-repo. No external service.
Why
A six-gram gate returned a FALSE NEGATIVE on material deposited five times. Until this exists, 'not previously deposited' is a claim the archive cannot make about its own work.
Approach
Concept-level overlap: distinctive-term sets, named-entity and formula matching, section-structure similarity — not n-grams.
Blocks
The blog compression audit; the Ω-Engine, Constitution and ASW seating runs; Gate 0 of EA-BLOG-RECOVERY-01.
Score the capture registry on the five-stage chain
PER-FACTORED
Requires
Python; the capture registry, in-repo. No external service. RUNNABLE OFFLINE.
Why
σ_PE factors as chunk:detach:rerank:compose:reattribute. 356 seated captures become PROFILES rather than scalars, and the four from tonight already separate cleanly.
Approach
Per-capture stage attribution against the registered rubric in #1536.
Blocks
Nothing. Pure upside on data already held.
Regenerate the blog index on a schedule
BLOG-INDEX-CADENCE
Requires
Python and a scheduler; the Blogger public feed. No credentials.
Why
2,939 posts indexed 2026-08-23; it is a snapshot with no refresh.
Approach
Wrap the start-index harvester; verify against openSearch:totalResults each run.
Blocks
The demonstrated-completeness rule depends on the index being current.
Regenerate api/counts.json
COUNTS-REGEN
Requires
repo write access.
Why
Still reads 1,520 deposits, generated 2026-08-21; the registry is at 1,537. Every fleet surface fetching it shows the stale figure.
Approach
Determine whether the pipeline should emit it; if so, wire it into the surfaces stage.
Blocks
Fleet-wide surface truth.
Per-surface projection in the capture gallery
PROJECTION-GROUPING
Requires
repo write access; the gallery builder is in-repo.
Why
One projection should carry one address's captures on ONE surface — that is what makes the longitudinal series legible. Ruled 2026-08-23; deferred as a rendering change touching existing records.
Approach
Group by (addr_id, surface) at render; addr_id keying unchanged.
Blocks
Longitudinal reading of any address across surfaces.
Split corpora originals into a companion repo
CORPORA-SPLIT
Requires
repo write access on two repositories; ~900 MB transfer.
Why
811 MB of 861 MB is original/; text/ is 50 MB. A 2 GB clone is a barrier to the external custodian the archive needs, and history cannot be rewritten because deposits pin commit SHAs.
Approach
text/, source.json, MANIFEST stay; original/ moves; manifests still hash the heavy repo. Split forward, never rewrite.
Blocks
Third-party harvest — the redundancy fix.
Assign resolution routes to 164 conditions
RESOLUTION-ROUTES
Requires
reading time — 164 conditions in datasets/prediction-ledger/conditions.jsonl.
Why
More than a third of registered falsification conditions do not state HOW they would resolve. A condition without a route is a gesture.