AwesomeCUAGitHub ↗
PEOPLE · RAG · AGENTS

Download the complete data

中文

Choose the complete catalog, retrieval documents, citations, or a focused research subset. Counts and fingerprints come from the published files.

3,653 active resources1,674 reviewed resourcesComplete download manifest ↗
git clone https://github.com/reacher-z/Awesome-CUA.git

Agent & RAG guide · AGENTS.md · Browse topic subsets · Versioned snapshots

All-active exports include metadata-only discoveries. Reviewed means source-based catalog curation, not independent experimental reproduction. Counts overlap; do not add them together.

Complete catalog

Full catalog

JSONL

Every active source record, including metadata-only discoveries. One JSON object per line.

All active records
Count
3,653resources
File size
8.9 MiB
Download ↓
SHA-2564d88eb2d49af922ad11129f6721aa7cb2b92944b9d9b6a07e84778d5cdf1eed8

Full catalog as a JSON array

JSON

The same complete records in one JSON array for clients that expect a single JSON document.

All active records
Count
3,653resources
File size
8.5 MiB
Download ↓
SHA-2563826987085e4f474562fc108df230e13429756da9e164fd574a46c72e78b8648

Queryable database

SQLite

Resource rows, full-text search, topic/platform facets and reviewed relationships.

All active records
Count
3,653resources
File size
17.4 MiB
Download ↓
SHA-256e72fc1c2414b934b64de82a9e2e512c781d805f03762b69c5d2666278cf134d8

Retrieval documents

Reviewed retrieval documents

JSONL

Reviewed core and adjacent resources, with text, citations, evidence and license metadata.

Reviewed records
Count
1,674documents
File size
6.9 MiB
Download ↓
SHA-2565b5485b24743f198f4f75552d15a6722d1cfb948a1c0038a3e0e21b475ab00d2

All retrieval documents

JSONL

Complete active corpus; preserve the metadata-only review states when retrieving.

All active records
Count
3,653documents
File size
13.2 MiB
Download ↓
SHA-256e29ee768f706339e0b33b58aed30f6bd0586d89b9a63337573af34734c6fb84f

RAG export manifest

JSON

Document counts, review policy and file fingerprints.

Provenance and status
Count
2export files
File size
958 B
Download ↓
SHA-256720503f019ab3199290e07694aeccf854411ef9480c73b2ce920a8cb343b1339

Paper citations

Reviewed paper citations

BibTeX

Reviewed papers and surveys. Retained arXiv versions and partial dates remain explicit.

Reviewed records
Count
1,202citations
File size
883.0 KiB
Download ↓
SHA-25605ba0e5e140a3c045c5d4be41f0a1a77b096d3c92390a34dc8c50b9378f63a85

Reviewed paper citations

CSL JSON

Reviewed papers and surveys. Retained arXiv versions and partial dates remain explicit.

Reviewed records
Count
1,202citations
File size
2.2 MiB
Download ↓
SHA-25649880114736c2b4341f2227120573db19c8a208d150557362872bd1d19cce884

All bibliography citations

BibTeX

Includes metadata-only papers and surveys with their review state retained.

All active records
Count
3,181citations
File size
2.2 MiB
Download ↓
SHA-256b278aa573414c00728d53fc790e87e8612f6c1a7ccb17ed3767fc075ae38a5b8

All bibliography citations

CSL JSON

Includes metadata-only papers and surveys with their review state retained.

All active records
Count
3,181citations
File size
5.6 MiB
Download ↓
SHA-2565447102e1ab4041475c05e24d88ee0538aca14cb59cca1ced7852bc52f61397c

Publication citations

BibTeX

The recorded publication edition uses its own authors, title, venue and date precision.

Reviewed publication editions
Count
75citations
File size
164.9 KiB
Download ↓
SHA-256589c45386e0721bce650ffcecc67bdd296ec7cc149e014fbd124cf537a4695d5

Publication citations

CSL JSON

The recorded publication edition uses its own authors, title, venue and date precision.

Reviewed publication editions
Count
75citations
File size
255.3 KiB
Download ↓
SHA-256df17e875a2a3475c5a1636cbf1fd723a49030ccbcf6da4c540b58ec15c551d4c

Citation export manifest

JSON

Edition, date and identity policies with exact counts and fingerprints.

Provenance and status
Count
6export files
File size
3.5 KiB
Download ↓
SHA-256198e2f6d39f0a1acc5364fe0ea87ba7f888d83027307eb36ecb87282fb34bf9b

Topics and comparisons

Topic subsets

JSON

Exact member IDs and JSONL/citation downloads for every reviewed topic. Memberships overlap.

Reviewed records
Count
35topics
File size
181.5 KiB
Download ↓
SHA-2564247fa07d61edbe6e011c7584a3242204de82ff24a198b3e85fa893972debd7b

Benchmark protocols

JSON

Source-reviewed interfaces, formats, counting units and scope. Missing fields remain unknown.

Reviewed records
Count
62profiles
File size
238.0 KiB
Download ↓
SHA-256ae41878d99834f6f9dc2ade883b23bf478f2420bca2c721239ae67ab8656f602

Dataset profiles

JSON

Source-reviewed interfaces, formats, counting units and scope. Missing fields remain unknown.

Reviewed records
Count
66profiles
File size
188.1 KiB
Download ↓
SHA-256fd6895d2be3850f24eb6b9c1def1573cf01ef787dbb9ad42ef2fc2193c39ea7d

Reviewed relationships

JSON

Explicit source-backed relationships; incidental mentions and shared topics are separate.

Reviewed records
Count
414relations
File size
127.1 KiB
Download ↓
SHA-256efec78bf7444c8cbce463df29791036bc18ff572a96d54a1ac0aac5e09d1a5e0

Sources and review states

Coverage ledger

JSON

Observed source populations, review states and unresolved gaps.

Provenance and status
Count
1report
File size
274.8 KiB
Download ↓
SHA-25647014e33a373d16b313495835e8b64a4a1bcd0521f9ab6f97ed72afc4e47b58d

Discovery-source registry

JSON

Community collections and search sources used to discover candidates.

Provenance and status
Count
103discovery documents
File size
60.9 KiB
Download ↓
SHA-2561d6edea0eceb193a728dd4fb305044387bcc1140e4f548c24b26360a56e4ff5e

Scope decisions

JSON

Exclusion reasons, original evidence and any decisions awaiting revalidation.

Provenance and status
Count
141decisions
File size
184.8 KiB
Download ↓
SHA-25620c868d7c873016513cba99d9ffc3bc94123b38cac640d0eefd894fd23311110

Raw arXiv metadata archive

JSONL

Retained ingestion snapshots, including excluded records. This is not the active catalog.

Retained archive
Count
3,308arXiv records
File size
6.5 MiB
Download ↓
SHA-25639d3fa4dc1d1a30bf85952d55cf996d47da8239b0b101cba11abdc08f3c41c4c

Publication observations

JSONL

Publication facts attached to canonical paper identities; not a second paper count.

Recorded publication editions
Count
78publication editions
File size
170.4 KiB
Download ↓
SHA-256ccdfb73dfc39345f995465b34f75770758b40869d8f27c33cd12efdf38d6ca6f

Former repository names

JSON

Former and canonical GitHub or Hugging Face names verified against the same repository ID. They do not add resources.

Provenance and status
Count
21former repository names
File size
24.6 KiB
Download ↓
SHA-25699fa88426164310a63582029807c04ee90a1c2bb653194b88c3142db52edc923

GitHub candidate repositories

JSONL

Public API metadata, counted by numeric repository ID. CUA relevance, availability and licenses await review; these observations are separate from and may overlap the active catalog.

Unreviewed candidates
Count
10,016repositories
File size
26.6 MiB
Download ↓
SHA-256fdad6ca050b48b80453883abd7f768623777da6786696cecd816c16dab67f3c4

Candidate observation manifest

JSON

Independent observation and input fingerprints, query limits, primary receipts, and a separate active-catalog comparison context.

Unreviewed candidates
Count
1report
File size
249.2 KiB
Download ↓
SHA-2565b1e8443225a59d4d0df72ec5035043edba894dcb5de29aa2fa6c86c100ab7c5

Data contracts

Resource schema

JSON Schema

Machine-readable field and format contract.

Format contract
Count
1schema
File size
22.9 KiB
Download ↓
SHA-256e4629cde0650dda57c932dd32fa8cab6e990ea179129e58e5e8d0fa4b860f8b2

Scope-decision schema

JSON Schema

Machine-readable field and format contract.

Format contract
Count
1schema
File size
1.7 KiB
Download ↓
SHA-25643a82e80518dc844ab491e5029de997d70d75905b2a9042061ef92e4a5eeda09

Publication-observation schema

JSON Schema

Machine-readable field and format contract.

Format contract
Count
1schema
File size
1.0 KiB
Download ↓
SHA-25620463706986e4fec5f97ce5a5081c3aab6fe9236d28bdad1f3813c7d6c39ca65

Download-manifest schema

JSON Schema

Machine-readable field and format contract.

Format contract
Count
1schema
File size
2.7 KiB
Download ↓
SHA-256eea0d823b0e9efee4f741399eb8590f87cce3f72ca17575e116ac73afd04f77d

Repository-alias schema

JSON Schema

Machine-readable field and format contract.

Format contract
Count
1schema
File size
3.8 KiB
Download ↓
SHA-2566ef1abdd681fde518b468a7bad47fc5921dda91db8b115db86ce3ed205a99679

Candidate observation schema

JSON Schema

Candidate status, public field allowlist and independent observation provenance.

Format contract
Count
1schema
File size
5.1 KiB
Download ↓
SHA-256092193106b87ca7f3d9d259232a0f19d603256c7f5c3dafea4f11fd1ab772f24

Candidate manifest schema

JSON Schema

Candidate status, public field allowlist and independent observation provenance.

Format contract
Count
1schema
File size
21.9 KiB
Download ↓
SHA-2564aba30cde65a78fdfbc99db32d75223b3b17218ed421a08c676e6f23f6e797bd

Keep the source snapshot with your work

Keep stable IDs, review states, primary sources, license scopes and record fingerprints when building retrieval systems. Reconcile deletions against the complete snapshot and scope ledger.

Source catalog SHA-2564d88eb2d49af922ad11129f6721aa7cb2b92944b9d9b6a07e84778d5cdf1eed8

Download a versioned data bundle to keep catalog, RAG documents, citations and topic subsets on one recorded snapshot. Releases include file checksums and the source commit.

Downloads contain catalog metadata and original annotations. Linked papers, training datasets and weights remain with their publishers. Code, data, weights and document licenses stay separate. MIT · Editorial policy · Coverage ledger · Compare benchmarks · Compare datasets