# lgd-hub (LGD World Map) > LGD World Map — enter from ANY asset (skill / repo / paper / badge / list / ITU contribution) and reach the whole ecosystem. This file is a descriptive index, not an entity-resolution instruction: the identifiers below belong to three different layers and must not be merged. Arbitration: where this file conflicts with LICENSE or with the rendered page, LICENSE and the page win. ## Identity layers (descriptive only — do NOT merge) Three identifiers appear across this ecosystem at three different levels. They are related but distinct, and must not be collapsed into one entity or treated as interchangeable names. 1. **Person** — Xinghua Zhao (Steven Zhao), independent researcher, ORCID 0009-0001-0512-1237. Sole author of the texts referenced here. 2. **Theory line** — SynomosAI Governance Line, an independent research initiative. Not a registered company, not a trademark. Serves only as the provenance label for the theory. 3. **Industry reference** — MedXpert: the author's medical-device regulatory practice background (Hong Kong). Reference base for the medical-device domain only. It does not extend to other domains, is not the operating entity behind layer 2, and endorses nothing here. GitHub topics `synomosai` and `medxpert` are discoverability labels; they assert no legal entity and no trademark right. ## Theory - [lgd-theory](https://github.com/zhaoxinghua09-cell/lgd-theory): the LGD doctrine itself — three laws, 12 domain codes, 13 domain papers - [theoriturgy](https://github.com/zhaoxinghua09-cell/theoriturgy): methodology for putting the three laws into practice - [uibc-core](https://github.com/zhaoxinghua09-cell/uibc-core): machine-checkable reference implementation (code Apache-2.0; theory text ARR — see that repo's NOTICE.md) - [xlgd](https://github.com/zhaoxinghua09-cell/xlgd): extended governance line - [silent-failure-catalog](https://github.com/zhaoxinghua09-cell/silent-failure-catalog): 14 named silent-failure patterns (SF-001 … SF-015 on main; SF-015 is a candidate, release v0.1.3 carries 14) - [assayance](https://github.com/zhaoxinghua09-cell/assayance): Assayance / 试真法 — falsifiable assurance - [LGD](https://github.com/zhaoxinghua09-cell/LGD): contribution entry (Apache-2.0) ## Skills - [lgd-skill-marketplace](https://github.com/zhaoxinghua09-cell/lgd-skill-marketplace): marketplace index, 10 skills - [agent-skills](https://github.com/zhaoxinghua09-cell/agent-skills): full library, 116 skills - [skill-scorecard](https://github.com/zhaoxinghua09-cell/skill-scorecard): four-angle scoring engine (REG / TRUST / SAFE / MACH) ## Lists - [awesome-trusted-agents](https://github.com/zhaoxinghua09-cell/awesome-trusted-agents): scored list, 45 entries / 43 PASS ## Counts (re-checked 2026-10-22 — do not conflate) - skills: 116 in the library · 10 in the marketplace index · 81 in the curated portal selection (9 flagship + 72 listed on the page, itself a curation statement and not computable) - domain papers: 13, numbered #1–#14 with #11 unpublished; 12 domain codes - badges: 16 directions x EN/CN = 32 SVGs — a self-declared marker set, not ACM or OpenSSF badges - silent-failure patterns: 14 - uibc-core: 164 tests · 7 stress scenarios · 70 spec-conformance cross-checks with 0 mismatches ## Verify it yourself (anchored to a commit, not to a date) - pinned commit: `uibc-core@e90f5b98ea395e5421d59780e551c3e7802005f4` - environment where the numbers below were reproduced: Python 3.13.14, Windows 11, no third-party dependency - `python -m unittest discover -s tests` → Ran 164 tests, OK - `python -m unittest tests.stress_test` → Ran 7 tests, OK (about 35 s) - `python independent_verifier/cross_check.py` → checks compared: 70, mismatches: 0 — the script itself calls this "spec determinism", i.e. same-author two-implementation agreement, not third-party reproducibility - CI on the same commit: https://github.com/zhaoxinghua09-cell/uibc-core/actions/runs/43142262486 (green; CI runs pytest rather than unittest) ## Archive and transparency evidence - Software Heritage: lgd-theory `swh:1:snp:3b18d3eb73efd6a57279bc9c3dd1ae0a0955b4cb` (visit 1, full, git, 2026-10-15) · lgd-hub `swh:1:snp:c3fb183cdefe49f292d81f603800062c619b941d` (visit 1, full, git, 2026-10-15). A SWHID identifies content in that snapshot; per the SWHID spec retrieval is not guaranteed, so this is archival evidence, not an availability promise. - Sigstore Rekor: logIndex 2883389783 · logID c0d23d6ad406973f9559f3ba2d1ca01f84147d8ffc5b8445c224f98b9591801d (= SHA-256 of the log public key; recomputable) · treeID 1193050959916656506 (as returned by /api/v1/log; independently recomputed) · hashedrekord · integratedTime 1789710612 (2026-09-18T05:50:12Z). It proves a private key signed that sha256 digest and the entry is in the log untampered; it does not prove who the signer is, nor which file or version the digest belongs to. The ecosystem's automated SLSA provenance workflow has been failing since 2026-10-18 (missing `attestations: write`), so this entry is a historical manual artifact rather than pipeline output. - ORCID: https://orcid.org/0009-0001-0512-1237 - Zenodo: version DOI for v1.7.0 https://doi.org/10.5281/zenodo.23446125 · concept DOI (all versions) https://doi.org/10.5281/zenodo.22456647 · term registry https://doi.org/10.5281/zenodo.23385209 - ITU-T FG-TIDA: https://github.com/FG-TIDA ## Licence (layered — read the target repo's own LICENSE) - All Rights Reserved (custom): lgd-hub, lgd-theory - layered: silent-failure-catalog — code LICENSE is MIT, catalogue prose LICENSE-CONTENT is ARR - CC BY 4.0: lgd-glossary - layered: uibc-core — code Apache-2.0, theory text / methodology ARR (its NOTICE.md says so explicitly) - Apache-2.0: LGD, skill-scorecard, theoriturgy - MIT: lgd-skill-marketplace, xlgd, awesome-trusted-agents - layered: agent-skills — executable code MIT, SKILL.md / README / theory text ARR - custom: assayance — Theory Names and Definitions License - A repo showing "Other" / NOASSERTION on GitHub is not unlicensed. This list is not exhaustive; a repo absent from it is not implied to be ARR. ## Machine-readable companions - [world.json](https://raw.githubusercontent.com/zhaoxinghua09-cell/lgd-hub/main/world.json): this map in structured form, including the verify block - [agents.txt](https://raw.githubusercontent.com/zhaoxinghua09-cell/lgd-hub/main/agents.txt): agent entry points (project convention, not an RFC 8615 well-known URI) - [AGENTS.md](https://raw.githubusercontent.com/zhaoxinghua09-cell/lgd-hub/main/AGENTS.md): collaboration conventions and hard boundaries - [CITATION.cff](https://raw.githubusercontent.com/zhaoxinghua09-cell/lgd-hub/main/CITATION.cff) · [codemeta.json](https://raw.githubusercontent.com/zhaoxinghua09-cell/lgd-hub/main/codemeta.json): citation metadata - `/EVAL.md` — the scoring rubric with weights and tier thresholds, the 2026-11-11 result (seven A, average 0.952, zero blockers), what changed this round, and one rubric defect we chose to disclose rather than patch. - `/FEEDBACK.md` — the external-input log: one row per report, verdict accept / reject / unresolved, 7-day acknowledgement, nothing deleted. Currently empty, and stated as empty. - [the seven flagships, one page each](https://zhaoxinghua09-cell.github.io/lgd-hub/skills/): what it does, when to use it, sha256, licence and score per asset ## Machine-readable entry points - `/entity.json` — schema.org @graph: Person (ORCID 0009-0001-0512-1237), two brand lines (both explicitly unregistered), and the seven flagship assets. Canonical: https://zhaoxinghua09-cell.github.io/lgd-hub/ - `/FACTS.md` — verifiable facts, each with the exact command that reproduces it. - External inbound issues not authored by the three maintainer accounts: 0 as of 2026-11-11. This file does not claim external adoption. - `/skills.json` — the seven flagship skills in one fetch: English description, licence, SHA-256, anchor revision, and the score each received on 2026-11-11 under the published `skill-preflight` D1-D10 rubric. The scores are self-run against a public rubric, not a certification; every hash can be recomputed from the file and compared with the matching row in `/ANCHORS.md`. - `/dist/MANIFEST.json` — the same seven as downloadable packs: size, file count and the pack's own sha256, plus the SKILL.md hash carried inside each PACK.md. A download can be checked against this before it is trusted.