Categories and coverage
Every category, whether or not anything has landed in it yet.
Plate 56 — water hyacinth with marbled crickets and frogs, Metamorphosis insectorum Surinamensium, Maria Sibylla Merian, 1705. Koninklijke Bibliotheek, The Hague (KW 1792 A 19). Public domain. Credits
On the counts
Thirteen categories: nine of this hub's own, and four shared lanes carried by both hubs of this survey. Counts state their denominators, and empty categories say so.
That statement is shared by both hubs of the survey, apart from this hub's own counts.
Category codes are stable; the numbering skips codes retired by earlier drafts (A5).
-
Literature & discovery
A1
Search, recall, and the instruments that calibrate both — including the campaign that assembled the survey's own recall-calibration toolkit.
1 of 5 campaigns have entries ingested here · 4 of 91 entries
Tasks
- Calibrate search & recall — 4 of 91 entries
- Journal & venue intelligence — 0 of 91 entries
- Mine research repositories & pipelines — 0 of 91 entries
- Run deep-research agents — 0 of 91 entries
- Living systematic reviews Planned — 0 of 91 entries
- Meta-analysis & evidence grading Planned — 0 of 91 entries
-
Data collection & curation
A2
Evidence capture, transcript harvesting, reference management, data governance — the layer where research data acquires provenance.
1 of 6 campaigns have entries ingested here · 8 of 91 entries
Tasks
- Capture & archive web evidence — 0 of 91 entries
- Govern research data — 8 of 91 entries
- Harvest transcripts — 0 of 91 entries
- Manage references — 0 of 91 entries
- Collector custody & resume Planned — 0 of 91 entries
-
Documents & data infrastructure
A3
Classification with humans in the loop, and the infrastructure that keeps research data queryable. Campaign rows are listed; entries are not yet ingested.
0 of 3 campaigns have entries ingested here · 0 of 91 entries
Tasks
- Build research data infrastructure — 0 of 91 entries
- Classify with humans in the loop — 0 of 91 entries
- Data lineage contracts Planned — 0 of 91 entries
-
Analysis & measurement
A4
Instruments and their validity: model judges, platform panels, simulated agent societies. The longitudinal measurement campaign here is the weekly design's methodological precedent.
1 of 5 campaigns have entries ingested here · 13 of 91 entries
Tasks
- Calibrate model judges — 0 of 91 entries
- Measure platforms longitudinally — 13 of 91 entries
- Simulate agent societies & markets — 0 of 91 entries
- Validate LLMs as instruments — 0 of 91 entries
- Preregistration & power Planned — 0 of 91 entries
- Psychometrics Planned — 0 of 91 entries
- Missing data Planned — 0 of 91 entries
-
Paper development & review
A6
Revision from feedback, claim and citation verification, hallucination reduction, provenance. The manuscript-revision campaign is ingested; its neighbors are listed with status shown.
1 of 5 campaigns have entries ingested here · 10 of 91 entries
Tasks
- Peer review & integrity machinery — 0 of 91 entries
- Provenance & scientific verification — 0 of 91 entries
- Reduce hallucination — 0 of 91 entries
- Revise manuscripts from feedback — 10 of 91 entries
- Verify claims & citations — 0 of 91 entries
- Statistical-checking audits Planned — 0 of 91 entries
- Venue & submission compliance Planned — 0 of 91 entries
-
Publishing & outreach
A7
One landed campaign: newsletters and outreach. Its ledger admitted no repositories, and the category's remaining tasks are still planned.
0 of 1 campaigns have entries ingested here · 0 of 91 entries
Tasks
- Newsletters & outreach — 0 of 91 entries
- Scholarly publishing pipelines Planned — 0 of 91 entries
-
Teaching
A8
Case engineering and course-document understanding. The teaching-case campaign's admitted set is ingested in full.
1 of 3 campaigns have entries ingested here · 4 of 91 entries
Tasks
- Engineer teaching cases — 4 of 91 entries
- Understand course documents — 0 of 91 entries
- Human-agent grading Planned — 0 of 91 entries
- Classroom experiments Planned — 0 of 91 entries
-
Theorization
A9
One campaign, ingested in full: tooling that scaffolds construct maps, mechanisms, and propositions. A thin shelf, kept open for the function it names.
1 of 1 campaigns have entries ingested here · 5 of 91 entries
-
Statistical computing
A10
Stata-centered evidence capture and reproducible result tables. Campaign rows are listed; entries are not yet ingested.
0 of 2 campaigns have entries ingested here · 0 of 91 entries
Tasks
- Statistical computing with Stata — 0 of 91 entries
- Causal-inference toolchains Planned — 0 of 91 entries
- Lab reproducibility Planned — 0 of 91 entries
-
Human-agent operations
S1 Shared lane
Human-agent operations: orchestration, runtime boundaries, memory, evaluation, security, and cost accounting. Token economics is a named sub-category.
1 of 20 campaigns have entries ingested here · 10 of 91 entries
Tasks
- Agent identity & interoperability — 0 of 91 entries
- Agent runtime & boundaries — 0 of 91 entries
- Design agent organizations — 0 of 91 entries
- Evaluate agents & skills — 0 of 91 entries
- Manage agent memory & context — 0 of 91 entries
- Operator & research consoles — 0 of 91 entries
- Orchestrate multi-agent workflows — 0 of 91 entries
- Package & port agent skills — 0 of 91 entries
- Run a solo-operator agent firm — 0 of 91 entries
- Secure the agent supply chain — 0 of 91 entries
- Token economics & AI cost — 10 of 91 entries
- Track tool & repo change — 0 of 91 entries
- Allocate tasks across humans & agents Planned — 0 of 91 entries
- Score orchestration Planned — 0 of 91 entries
- MCP/tool interoperability Planned — 0 of 91 entries
-
Tools & document processing
S2 Shared lane
Documents in, structure out: parsing tooling measured under real workloads, with benchmarks kept distinct from parsers.
1 of 2 campaigns have entries ingested here · 12 of 91 entries
Tasks
- Parse PDFs & documents — 12 of 91 entries
- Multi-format document intelligence Planned — 0 of 91 entries
-
Writing & editing
S3 Shared lane
Words are cheap to generate now, and it shows. Tooling that checks prose against its sources — editing, cleanup, translation.
1 of 10 campaigns have entries ingested here · 11 of 91 entries
Tasks
- Academic writing stack — 0 of 91 entries
- De-AI & polish prose — 0 of 91 entries
- Evidence-grounded editing — 11 of 91 entries
- Redline contracts & documents — 0 of 91 entries
- Translate EN<->ZH — 0 of 91 entries
- Write business documents — 0 of 91 entries
-
Presentation & visuals
S4 Shared lane
Slides, figures, infographics — the tooling that turns work into something an audience can follow, with editability and attestation treated as evidence.
1 of 7 campaigns have entries ingested here · 14 of 91 entries
Tasks
- Figures & data visualization — 0 of 91 entries
- Make infographics — 0 of 91 entries
- Make slides, decks & talks — 14 of 91 entries
- Evidence-to-visual traceability Planned — 0 of 91 entries
Absent by design
What this site does not do, and why.
Three things this site refuses, and three it has not performed.
| Area | State | Detail |
|---|---|---|
| Rankings | Refused | No ordering on this site encodes quality. Category and name only. |
| Repo-authored text | Refused | No README excerpts, project descriptions, or logos appear anywhere. Metadata facts only. |
| Identity lists | Refused | Counts only. No stargazer, watcher, or contributor identities, ever. |
| Peer review | Not performed | Nothing here has been peer-reviewed by us, and presence implies no such thing. |
| Replication | Not performed | We did not run, rebuild, or reproduce any repository's claims. |
| Verification | Not performed | Functional verification was not performed. The survey records what ledgers admitted, not what works. |